💻

Computer Science

18 Projects

Project #1 Group A-A

GoFix- A Smart App for Home Services

Supervisor Dr. Yaseen ul Haq & Miss Rabia Sana

Abstract

Home service systems are becoming common as people look for quick and reliable ways to connect with electricians, plumbers, and other technicians. However, most of the available systems are limited because they work on only one platform, provide very basic features, and do not offer smart support for users. Because of these issues, people find the system hard to use. It becomes difficult to quickly find the right technician, communication is not clear, GPS tracking does not work properly, and users cannot fully trust the service providers. Previous work in this field has tried to solve these issues but usually stayed at a simple level, focusing only on service listings and direct bookings. The proposed smart GoFix system aims to solve these challenges by offering a smart solution. It supports two main roles: customers and technicians. Customers can write a short description of their problem, attach an image, enter an estimated budget, and provide their location. Using GPS, the system automatically finds technicians available in the same city. Technicians can then place bids, and once the customer accepts an offer, both sides can communicate through direct phone calls or in-app chat. A clear and welled fined cancellation process is also provided, where users must state valid reason before canceling, ensuring fairness and transparency for both parties. The Smart GoFix system also incorporates AI-powered features such as smart price estimation, which facilitates fair cost agreements between customers and technicians, and an AI assistant for technicians that allows them to report their problems and receive immediate guidance and solutions. An AI-powered, transparent platform for connecting customers and technicians is provided by the GoFix system. It ensures trust, fairness, and reliability through smart features, and effective admin oversight.

Authors

NameRegistration No.
Shahzaib Hassan2022-CS-505
Muhammad Naeem Atif2022-CS-555
Mirza Sufiyan Ahmad2022-CS-513
Sadia Hamid2022-CS-537
Project #2 Group A-B

Tabib Online - AI Enabled HealthCare Appointment Platform

Supervisor Miss Rabia Sana

Abstract

Tabib Online is an AI-powered healthcare platform designed to revolutionize medical consultations in Pakistan by connecting patients with verified doctors through intelligent symptom analysis and automated appointment booking. The system features an AI agentic bot trained on Pakistani medical literature that provides culturally relevant health guidance, automatically escalates serious cases to specialists, and handles comprehensive healthcare navigation. Core functionality includes automated doctor verification through PMDC integration, multi-modal consultations (video, audio, in-clinic), and secure digital health record management. The platform addresses healthcare accessibility challenges in underserved regions while maintaining strict data security standards and providing analytics dashboards for healthcare providers, ultimately improving healthcare outcomes through AI-driven automation and user-centric design.

Authors

NameRegistration No.
Muhammad Abdullah Zahid2022-CS-525
Aiza Eman2022-CS-563
Asim Shahzad2022-CS-547
Project #3 Group A-C

Revogue 3D: Virtual Try-on Solution

Supervisor Miss Rabia Sana & Miss Rabia Zafar

Abstract

The “Revogue 3D: Virtual Try-on System” is a web application designed to enhance shopping experience for customers. It allows users to visualize how different clothes look on them without physically wearing them. The system utilizes Augmented Reality and Pose detection technologies to overlay virtual clothes onto a user’s image in real time, helping users make better purchasing decisions and reducing product return rates. In addition, the system 3D aims to bridge the gap of online shopping. This platform offers an interactive interface to allow users to browse various outfits, view and adjust sizes and explore different styles to find the most suitable options according to their choice. By combining Computer vision and AR-based rendering, the system delivers a user-friendly solution that enhances customer satisfaction and provides an interactive, realistic virtual try-on experience for customers.

Authors

NameRegistration No.
Nazima Latif2022-CS-501
Muhammad Mudassar Younas2022-CS-535
Ghanwa Tahir2022-CS-570
Rubia Ali Toor2022-CS-553
Project #4 Group A-D

Zabaan e Kisaan-AI and Urdu Voice based Assistant for Farmer

Supervisor Miss Rabia Sana & Dr. Yaseen ul Haq

Abstract

This document presents the design and development of Zabaan-E-Kissan (AI & Voice-Based), an intelligent agricultural assistance system aimed at empowering farmers in rural and underserved regions through accessible and technology-driven solutions. The system addresses critical challenges faced by farmers, including limited access to expert knowledge, unpredictable weather conditions, pest infestations, and fluctuating market dynamics. The proposed solution leverages Artificial Intelligence (AI) and Natural Language Processing (NLP) to enable seamless interaction through voice-based communication in local languages such as Urdu, making it highly accessible for users with low literacy levels. The platform is developed using cross-platform mobile technologies, allowing farmers to interact with the system via voice commands and receive responses in both textual and spoken formats. A key feature of the system is AI-powered crop disease and pest detection, which utilizes image processing and machine learning models to analyze crop images and provide accurate diagnoses along with recommended treatments. Additionally, the system integrates irrigation guidance, and market price updates, offering personalized and region-specific agricultural advice. The system aims to bridge the gap between modern technology and traditional farming practices by providing timely, reliable, and localized information. By enhancing decision-making capabilities, improving crop management, and increasing productivity, Zabaan-E-Kissan contributes to sustainable agriculture and improved farmer livelihoods.

Authors

NameRegistration No.
Abdulahad Hussain2022-CS-549
Zoha Afzal2022-CS-531
Mehreen Farooq2022-CS-511
Project #5 Group A-E

Export360 ERP System

Supervisor Dr. Yaseen ul Haq & Miss Rabia Sana

Abstract

Export 360 ERP is a cross-platform, multilingual (Urdu and English), and cost-effective enterprise resource planning solution designed to digitize and modernize the export industry of Pakistan. The system ensures accessibility across desktop and mobile devices while providing a user-friendly interface tailored to local business needs, making it suitable for small and medium sized exporters. Pakistan’s economy heavily relies on the agricultural sector, where products such as rice, wheat, corn, and turmeric contribute significantly to exports. Among these, the rice export industry plays a crucial role; however, many rice mills still depend on manual record-keeping and fragmented systems for managing operations. This results in inefficiencies, data inconsistency, and limited business insights. To overcome these challenges, Export 360 ERP provides a comprehensive solution specifically focused on rice mill management. The system introduces an automated inventory management module that tracks the complete lifecycle of goods, including raw materials, processing stages, by-products, and waste. It also includes key operational features such as automated gate pass generation, loan management, advance salary tracking, and installment handling. Furthermore, the system enhances financial operations through integrated banking features and a secure digital cash drawer for managing transactions. A smart dashboard enables users to monitor daily and weekly activities while offering graphical insights into sales, purchases, and profit-loss patterns for better decision-making. By integrating all core business processes into a single platform, Export 360 ERP reduces manual errors, improves efficiency, and provides real-time data access. The system aims to support the digital transformation of Pakistan’s export sector by delivering a localized, scalable, and efficient ERP solution.

Authors

NameRegistration No.
Faisal Faiz2022-CS-523
Maha Butt2022-CS-527
Laraib Ishtiaq2022-CS-508
Project #6 Group A-F

Automating Cyber Threat Hunting using NLP and MAS

Supervisor Miss Fatima Shahzadi & Dr. Sadia Tariq

Abstract

The fast development of cyber threats, from advanced persistent threats (APTs) to zero-day exploits and artificial intelligence (AI)-driven attacks, has made apparent a number of shortcomings in current approaches based on rule-based and signature-based security. Therefore, the proposed method aims to combine natural language processing (NLP) and multi-agent systems (MAS) to provide a more efficient solution to detect cyber threats. Specifically, the first part of the system will employ NLP tools to analyze large volumes of unstructured data in the form of reports on cyber threats, disclosure of software vulnerabilities, social media messages, and other data on the dark web. These include tasks such as text preprocessing (tokenization, removing stop-words, stemming), extracting features using various techniques (for example, TF-IDF or word embeddings) and classifying the data through machine learning and deep learning algorithms to determine the level of threat. In doing so, the algorithm can identify the emerging cyber threats by analyzing textual information. The second part of the algorithm analyzes structured data with a multi-agent system, whereby autonomous agents will be equipped with certain roles, including monitoring of network traffic, analyzing system logs, anomaly detection, etc. The collaboration and communication between these agents allow for exchange of information, use of reasoning skills, and decision-making. For instance, one agent may notice some irregularities in the logins, whereas another notices unusual activities with files, and both agents may conclude that an attack is underway. As a result of the integration of both modules, the system gains the ability to process unstructured data through contextual analysis and structured data through operational analysis. This allows for the creation of a unified security defense system that not only recognizes all known attacks but also adapts to new and innovative attacks, unlike traditional security systems. Moreover, the system may include a feedback loop that uses the results of the incidents detected to train the NLP model and agents’ behavior, which improves its performance over time. The system also provides for scalability since several agents may be used to accomplish different tasks. The intelligent decision making ensures rapid response to incidents and minimal human involvement. Moreover, reinforcement learning can be incorporated into the system to improve the strategies of the agent through past experience. All in all, this integrated solution not only improves the ability of detecting threats but also leads to developing an evolving cyber security infrastructure.

Authors

NameRegistration No.
Maham Liaqat2022-CS-503
Syed Subtain Ali2022-CS-529
Ali Haider2022-CS-538
Muhammad Murtaza (Section B)CD-CS-01
Project #7 Group A-G

AI Powered Surveillance: Detecting Suspicious Chats, Crowd Activities and Heatmap based anomalies+D9

Supervisor Dr. Sadia Tariq

Abstract

The rapid expansion of digital communication platforms and urban populations has significantly increased the complexity of public safety management. Traditional surveillance systems, which rely on manual monitoring of video feeds and textual data, are increasingly inadequate due to inefficiency, susceptibility to human error, and limited real-time processing capabilities. These limitations hinder timely threat detection, particularly in high-density environments where continuous monitoring is required. The challenge is further intensified in developing regions such as Pakistan, where a substantial portion of communication occurs in Roman Urdu — a non-standardized language characterized by inconsistent spelling, grammar, and syntax. This linguistic variability significantly reduces the effectiveness of conventional Natural Language Processing (NLP) systems in accurately interpreting and analyzing potentially suspicious or harmful content, leading to missed threats and reduced system reliability. To address these challenges, this thesis proposes an AI-powered Multi-Modal Surveillance System that integrates Natural Language Processing and Computer Vision to enable automated and real-time threat detection. The NLP module is designed to process and classify Roman Urdu text into suspicious and non-suspicious categories using a hybrid architecture combining rule-based keyword detection with MiniLM sentence embeddings and Logistic Regression classification. The Computer Vision module analyzes live video streams to identify abnormal crowd behaviors such as violence, panic, and overcrowding using a YOLOv8-based classification model trained on a combined dataset of 17,749 training images and 2,310 validation images sourced from the Kaggle Real Life Violence dataset and the AIRTLab Violence Detection dataset. A key contribution of this research is the development of a fusion mechanism that combines insights from both textual and visual modalities, thereby enhancing detection accuracy and minimizing false positives. Additionally, the system incorporates advanced visualization techniques, including Gaussian-based heatmaps and optical flow analysis, to monitor crowd density and detect anomalies in real time. Experimental evaluation demonstrates strong performance, with the NLP module achieving 97.3% accuracy and the Computer Vision module achieving 97.2% accuracy for fighting detection, both surpassing all fourteen prior works reviewed in the literature. The integrated multi-modal system achieves 97.5% overall accuracy, representing a 0.2% improvement over the NLP module and a 0.3% improvement over the Computer Vision module individually, demonstrating the effectiveness of feature-level fusion across both modalities. The proposed solution is designed for scalability and real-time deployment on cloud infrastructure (AWS EC2) and edge computing platforms, delivering alerts with a latency of under 0.5 seconds, making it a practical and efficient approach for real-world surveillance applications. The proposed multi-modal fusion model achieves an accuracy of 97.5%, showing improvement over the standalone NLP model (97.3%) and the standalone Computer Vision model (97.2%). This demonstrates the effectiveness of integrating textual and visual modalities, resulting in enhanced detection reliability and reduced false positive rates across diverse surveillance scenarios.

Authors

NameRegistration No.
Muhammed Rameez2022-CS-543
Muneeba Tehreem2022-CS-544
Abubaker Saddique2022-CS-554
Ali Haider2022-CS-509
Project #8 Group A-H

EcoGuard: A Comprehensive Deep Learning Framework for Environmental and Maritime Surveillance Using Satellite Imagery

Supervisor Dr. Sadia Tariq

Abstract

Environmental degradation and maritime security threats represent two of the most pressing challenges of the twenty-first century, demanding scalable and automated surveillance solutions. This thesis introduces EcoGuard, a comprehensive web-based deep learning platform engineered for dual-domain satellite surveillance: terrestrial environmental monitoring for illegal deforestation detection, and maritime domain awareness for unauthorized small vessel detection. The system leverages two custom-built YOLOv8 object detection architectures, trained entirely from scratch on multi-source satellite imagery, deliberately avoiding transfer learning from terrestrial datasets to maximize performance on aerial perspectives. For the deforestation module, optical imagery from Sentinel2, Landsat 8/9, and Google Earth was used to train the model to detect five key land-cover classes: trees, tin-shade structures, water bodies, houses, and buildings. Semi-supervised learning with pseudo-labeling was employed to leverage both labeled and unlabeled data collected dynamically from diverse global locations, achieving an overall detection accuracy of approximately 94%. The maritime module exclusively employs Synthetic Aperture Radar (SAR) imagery from Sentinel-1 and commercial SAR platforms to overcome the cloudcover limitations of optical sensors. The custom YOLOv8 architecture was tuned specifically for SAR speckle noise profiles and small vessel signatures, achieving approximately 94% detection accuracy across diverse sea conditions. EcoGuard is deployed as a responsive web application with a React and CSS frontend featuring three main pages: Home, Deforestation Detection, and Vessel Detection. Both detection interfaces support user-adjustable confidence thresholds, display the total number of detections per image with individual confidence percentages, and algorithmically compute the physical dimensions — length, width, and estimated height — of each detected object using satellite metadata and geometric mensuration. Users may export all detection results and dimensional data as downloadable PDF reports. The platform also exposes a secure public REST API for external integration, and incorporates educational article links enabling users to learn about the underlying technologies.

Authors

NameRegistration No.
Ihtisham Ali2022-CS-507
Abaidur E Rehman2022-CS-565
Abdul Sami (Section B)2022-CS-560
Muhammed Tahir2022-CS-517
Project #9 Group A-I

Dysgraphia Detection

Supervisor Dr. Iqra Muneer

Abstract

Dysgraphia is a neurological handwriting disorder estimated to affect between 5% and 20%of school-age children globally. Despite its prevalence, it is systematically underdiagnosed, primarily because clinical assessment depends on scarce specialist professionals and expensive standardized instruments. The existing body of computational dysgraphia research is confined exclusively to Latin-script, English language data, leaving Urdu, a right-to-left Nastaliq script language with over 230 million speakers, entirely unaddressed. This thesis presents the design, implementation, and evaluation of the first bilingual dysgraphia detection system for school-age children in Pakistan, supporting both English and Urdu handwriting. Four EfficientNetB0 + Logistic Regression model configurations are developed and evaluated: English sentence-level (247 samples), English word-level (1,086 samples), Urdu sentence-level (243 samples), and Urdu word-level (1,000 samples). A custom dataset of 2,576 clinically-labelled handwriting samples was collected from Pakistani school-age children aged 8–16, the first structured, labelled Urdu dysgraphia corpus in the literature. The uniform pipeline applies frozen EfficientNetB0 features (1,280-dimensional vectors) as input to a Logistic Regression classifier, evaluated under Stratified 10-Fold Cross-Validation using six metrics: Accuracy, Precision, Recall, F1-Score, Specificity, and ROC-AUC. Class imbalance in the Urdu sentence-level dataset is addressed through SMOTE and class-weighted training. The experimental results provide the first empirical evidence on the script agnostic properties of EfficientNetB0 features across Latin and Nastaliq scripts, a methodologically generalizable finding with implications for multilingual handwriting analysis beyond the specific context of dysgraphia detection. The trained models are deployed as a modular web application (React.js / Node.js / AWS) that returns diagnostic reports within a 30-second window, making preliminary dysgraphia screening accessible to educators and caregivers without specialist training.

Authors

NameRegistration No.
Faaiz Mahmood2022-CS-515
Awais Amjad2022-CS-541
Sufyan Rafaqat2022-CS-533
Fahad Ali (Section B)2022-CS-524
Project #10 Group B-A

ZAKOOTA – A Layer Hiring App

Supervisor Miss Madiha Maqbool & Miss Farva

Abstract

ZAKOOTA is a mobile-first digital freelance marketplace that connects clients with verified lawyers in Pakistan through a secure and transparent platform. Built using Flutter (Dart) for cross-platform support and Firebase as the cloud backend, the application allows clients to post legal tasks, search and filter verified lawyers by specialization, book consultations, communicate via real-time in-app chat, and process payments through JazzCash and Easypaisa. A key innovation of ZAKOOTA is its built-in AI Legal Chatbot, Zing AI, powered by the Groq API (Llama 3.1). Zing AI assists users by explaining their legal issues in plain English, identifying the relevant area of Pakistani law (family, criminal, tenancy, property, corporate, etc.), and guiding them toward verified lawyers on the platform. The chatbot is enhanced with a curated Pakistani legal knowledge base covering the Pakistan Penal Code, Muslim Family Laws Ordinance, PECA 2016, and more. The platform addresses a clear gap in the Pakistani legal market — no existing solution combines AI-powered legal guidance, Pakistan Bar Council credential verification, gig-based lawyer listings, real-time chat, and local payment integration in a single application. User acceptance testing with 8 participants yielded an average satisfaction score of 4.2/5, with 88% chatbot accuracy across 50 Pakistani legal queries.

Authors

NameRegistration No.
Inza Iqbal2022-CS-502
Rumaisa Aman2022-CS-518
Faisal Khalid2022-CS-528
Waleed Bin Tahir2022-CS-546
Project #11 Group B-B

HIREFLOW - AI-Powered Interview Platform for Bulk Hiring

Supervisor Miss Fatima Shahzadi

Abstract

HireFlow is an AI-powered interview platform designed to automate and streamline the recruitment process for bulk hiring. HireFlow aims to revolutionize the recruitment process by automating interviews for mass hiring through artificial intelligence, leveraging natural language processing and computer vision technologies to conduct real-time interviews. The system generates interview questions using Google Gemini model, ensuring relevance to the job role and requirements. Candidates can respond to coding questions in the Code Editor while verbal answers are transcribed using Speech-to-Text service and fed into the response pane. Both types of responses are evaluated alongside non-verbal cues such as facial expressions and eye contact. Computer vision tools like MediaPipe are used to analyze facial expressions and detect anomalies and any cheating attempts during the interview. The platform features two distinct modules: HR Dashboard for recruitment management and Interviewee Dashboard for job seekers. Each candidate receives a unique interview link, allowing them to complete their assessment remotely. The system evaluates responses, flags suspicious behavior, and generates comprehensive performance reports with scoring and rankings. The Literature review surveys more than ten distinct parallel approaches to automated recruitment, spanning mock interviews, commercial video platforms, fairness, ethics, proctoring, question generation, and related topics; these lines of work are often pursued in isolation. HireFlow differs by integrating them on a single platform: end-to-end job and applicant management, interview conduct, monitoring, and AI-assisted evaluation for many concurrent candidates. The intended impact is higher throughput, more consistent screening, lower operational cost while providing a secure, scalable, and automated solution for modern human resource management.

Authors

NameRegistration No.
Muhammad Awais Mushtaq2022-CS-530
Usman Younas2022-CS-534
Jabir bin Yaqub2022-CS-552
Project #12 Group B-C

AI Based Smart House Construction Estimator Dashboard

Supervisor Dr. Yaseen ul Haq & Miss Rabia Sana

Abstract

The construction industry often faces challenges such as inefficient cost estimation, lack of transparency in contractor selection, poor communication between clients and contractors, and inability to track real-time project progress. Traditional methods of house construction planning are time-consuming, error-prone, and heavily dependent on manual expertise. Current AI-based approaches attempt to predict costs without contractor expertise, leading to inaccurate estimates. To address these issues, this project proposes an AI-Based Smart House Contractor Bidding System, a web-based intelligent platform that integrates artificial intelligence, modern web technologies, contractor expertise, and real-time communication to streamline the entire construction planning and bidding process. The system allows users to input their plot dimensions and construction requirements, after which an AI-powered computer vision model generates accurate 2D architectural layouts and preliminary design suggestions. Contractors can register on the platform, build verified profiles, and submit competitive bids based on the AI-generated designs and detailed project specifications. As the project progresses, contractors provide daily progress reports that continuously refine cost estimations, ensuring accurate and transparent project tracking. Clients can compare bids, evaluate contractor ratings, monitor real-time progress, and receive accurate cost breakdowns based on actual contractor expertise and daily project updates. The platform is developed using the MERN stack (MongoDB, Express.js, React.js, Node.js) with Python backend integration for AI/ML operations, ensuring scalability and responsiveness. Machine learning models using computer vision are integrated for 2D map generation and layout optimization, while WebSocket technology enables real-time communication between stakeholders. The system includes real-time progress tracking with daily updates, secure authentication using JWT, an admin panel for system monitoring, contractor verification modules, and a comprehensive rating mechanism based on actual project delivery performance. This solution aims to reduce construction planning time through automated 2D design generation, improve cost accuracy through contractor expertise and daily progress tracking, and enhance transparency in contractor-client interactions. The proposed system contributes to the digital transformation of the construction industry by introducing an intelligent, efficient, and collaborative platform for smart house planning that leverages both AI capabilities and human contractor expertise.

Authors

NameRegistration No.
Huzaifa Arif2022-CS-556
Muhammad Abdullah2022-CS-536
Ameer Hamza2022-CS-562
Project #13 Group B-D

Unified ASD Screening, Diagnosis, Therapy, And Progress Tracking System

Supervisor Dr. Sadia Tariq

Abstract

Autism Spectrum Disorder (ASD) is a neurodevelopmental condition that affects communication, behavior, and social interaction. While early screening and timely intervention significantly improve developmental outcomes, access to structured, integrated, and technology-supported care remains limited, particularly in under resourced healthcare settings. To address this gap, this thesis presents the design, development, and evaluation ofAutismCare: a comprehensive, web-based platform for early ASD screening, diagnosis, therapy management, laboratory reporting, and developmental tracking for children under four years of age. The system was developed using the MERN stack (MongoDB, Express.js, React.js, and Node.js) and integrates a multi-modal screening approach combining standardized questionnaires with AI-based facial morphology analysis to identify autistic markers. Key functional requirements include parent registration, AI-integrated screening, appointment scheduling, clinician diagnosis workflows, therapy progress tracking, laboratory report handling, and role-based access control. Non-functional requirements emphasize performance, security, safety, and usability. The system was evaluated through unit, integration, and user acceptance testing, demonstrating compliance with defined benchmarks. Achieving 95% response time of less than 3 seconds for dashboard loading and completing complex AI inference in almost 10 seconds, the platform significantly optimizes clinical triaging. Statistical performance evaluations indicate a system uptime of 99.5%, ensuring reliable access for parents, clinicians, and therapists. By automating manual reporting and centralizing data flow, AutismCare demonstrates a 60% improvement in communication efficiency between stakeholders, facilitating timely diagnosis and continuous developmental monitoring. The results confirm that AutismCare provides a se cure, scalable, and user-centered solution that bridges the gap between parents and healthcare professionals, reduces diagnostic delays, and enables continuous, data-driven monitoring of early childhood development.

Authors

NameRegistration No.
Hassam Akram2022-CS-542
Aman Fatima2022-CS-506
Rabia2022-CS-561
Hafsa Mudassar2022-CS-569
Project #14 Group B-E

DOORBOT - Smart Door Automation System

Supervisor Miss Fatima Shahzadi

Abstract

DoorBot is an Internet of Things (IoT) based smart door automation and remote access control system designed to enhance home security and convenience. The system integrates a physical door unit, a cloud backend, a relay server, and a cross platform mobile application into a unified real time access management solution. The hardware subsystem is built around an AI Thinker Espressif Systems 32-bit Microcontroller (ESP32)32 Camera Module (CAM) module that captures live Joint Photographic Experts Group (JPEG) video frames at five frames per second and transmits them to a cloud relay server over Hypertext Transfer Protocol Secure (HTTPS). A homeowner can view this live feed from anywhere in the world through a Flutter based Android application, remotely unlock the door with a single tap, and receive instant push notifications when the physical doorbell button is pressed. All door events and doorbell activity are logged to Firebase Realtime Database (RTDB) and presented to the user through a date filterable history screen within the mobile application. A separate offline subsystem, built around an Arduino microcontroller and an R305 optical fingerprint sensor, provides physical access control that operates independently of network connectivity. Enrolled family members can unlock the door through fingerprint scanning. The backend architecture employs a Node.js relay server hosted on Render.com, which solves the problem of direct device accessibility behind residential Network Address Translation (NAT) routers. Firebase Cloud Messaging (FCM) push notifications are dispatched server side through the Firebase Admin Software Development Kit (SDK), ensuring delivery in all application lifecycle states including foreground, background, and terminated. The mobile application follows the Model-ViewViewModel (MVVM) architectural pattern using the Provider state management library. Security is enforced through Firebase Authentication with mandatory email verification, per user scoped RTDB security rules, Transport Layer Security (TLS) encryption for all network communication, and Application Programming Interface (API) key authentication between all system components. Functional and nonfunctional testing confirms that remote door unlock commands are executed within two seconds, push notifications are delivered within three seconds of a doorbell press, and live stream latency remains below 500 milliseconds under stable Wi Fi conditions. DoorBot demonstrates that a comprehensive, secure, and affordable smart door solution can be constructed using commodity hardware and opensource software frameworks.

Authors

NameRegistration No.
Muhammad Husnain Nawaz2022-CS-512
Tnazeel Ashraf2022-CS-568
Bilal Khan2022-CS-516
Muhammad Bilal Ahmad2022-CS-521
Project #15 Group B-F

NOCODENEXUS: No Code Data Preprocessing Tool

Supervisor Miss Rabia Sana

Abstract

NoCodeNexus is a no-code, web-based data preprocessing platform designed to help users clean and prepare datasets without programming knowledge. The system provides an easy-to-use interface for performing preprocessing tasks such as handling missing values, removing duplicates, encoding categorical data, feature scaling, feature engineering, and Principal Component Analysis (PCA). Built using the MERN stack with Python integration, the platform simplifies complex data preprocessing operations and improves productivity for students, researchers, and non-technical users by providing an efficient and interactive environment for data preparation.

Authors

NameRegistration No.
Dular Fatima2022-CS-519
Syed Huzafa2022-CS-520
Noman2022-CS-564
Project #16 Group B-G

TRANS LINGO - Real-Time Speech & Gesture Translation App for Communication

Supervisor Miss Madiha Maqbool & Miss Habiba Bashir

Abstract

Communication barriers between the Deaf and Hard-of-Hearing community and the general population pose a significant societal challenge, particularly for localized sign languages like Pakistan Sign Language (PSL). This thesis presents the development of TransLingo, a comprehensive, multi-modal sign language recognition and translation system specifically designed to bridge the communication gap for PSL users in Pakistan. The system recognizes five distinct categories: English Alphabets, English Adjectives, Urdu Alphabets, Urdu Words, and dynamic sequence-based Appliance signs. Utilizing Google's MediaPipe framework for robust, environment-invariant spatial feature extraction, the system employs a hybrid algorithmic approach that applies the most computationally efficient machine learning algorithm for each specific task. Support Vector Machines (SVM) with Radial Basis Function kernels are utilized for static gestures, Deep Dense Neural Networks for complex multi-hand static signs, and a novel Multi-Modal Bidirectional Long Short-Term Memory (Bi-LSTM) network coupled with Multi-Head Temporal Attention for dynamic video-based sequences. The Multi-Modal Bi-LSTM architecture processes 144-dimensional feature vectors extracted from both hands and upper body pose across 30-frame temporal sequences. Modality-specific encoders independently process left hand (63 dimensions), right hand (63 dimensions), and upper body pose (18 dimensions) features before fusion. The architecture achieved exceptional performance, demonstrating the effectiveness of the multi-modal approach for sign language recognition. The inference engine operates as a stateless Flask REST API, offering seamless real time integration with both a web-based interface and a cross-platform Flutter mobile application. The system provides instantaneous text-to-speech (TTS) auditory feedback, functioning as a practical, low-cost solution for communication inclusion. Advanced training methodologies including label smoothing, mixup regularization, cosine annealing with linear warmup, and gradient clipping ensure robust model performance. This research contributes to the field of assistive technology by providing the first comprehensive multi-modal PSL recognition system, creating custom datasets for under resourced sign language research, and demonstrating that lightweight, CPU-based deployment can achieve real-time performance suitable for practical communication scenarios.

Authors

NameRegistration No.
Muhammad Hassan (Section A)2021-CS-544
Muhammad Zameer ul Hassan2022-CS-540
Haris Khan2022-CS-557
Project #17 Group B-H

ZOOMIGO - The Ride Hailing App

Supervisor Miss Rabia Sana & Dr. Yaseen ul Haq

Abstract

The transportation landscape has transformed dramatically through digital innovation, with ride-hailing platforms emerging as a pivotal solution for urban mobility. This thesis delineates the design and development of Zoomigoo, a comprehensive ride-hailing ecosystem connecting commuters with nearby drivers via mobile applications. Zoomigoo was conceived to tackle the challenges faced by local commuters in accessing dependable, economical, and secure transportation. The platform encompasses three core mod- ules: a Passenger Application for ride booking and trip monitoring, a Driver Application for request acceptance and earnings management, and an Administrative Panel for platform oversight and user administration. The development harnessed contemporary technologies: Flutter and Dart for cross-platform mobile development, Node.js with Express for backend infrastructure, and MongoDB for data persistence. Integration with Google Maps APIs facilitated real-time location tracking and route optimization, while local digital payment systems including Easypaisa and JazzCash enabled secure transactions. Notable system features include OTP-verified user registration, live ride tracking, dynamic fare computation, multiple payment modalities, comprehensive driver verification, and robust administrative tools. The application supports bilingual functionality (English and Urdu) to maximize accessibility. The system underwent rigorous functional, performance, and user acceptance testing. Results demonstrated its capability to handle concurrent users, process ride requests efficiently, and deliver a seamless experience. Zoomigoo successfully fulfills the requirements of a modern ride-hailing platform while catering to local market needs.

Authors

NameRegistration No.
Maasma Zari2022-CS-504
Sakhra Zahid2022-CS-522
Muhammad Ali2022-CS-548
Project #18 Group B-I

Attention Aware ppt Reader App

Supervisor Dr. Sadia Tariq

Abstract

This thesis presents an intelligent Attention-Aware Document Reader that enables hands-free document interaction using eye-tracking, gesture recognition, and attention monitoring techniques. The system integrates gaze-based navigation, gesture controlled zooming, and real-time attention detection to improve user interaction and reading efficiency. The proposed framework uses webcam-based gaze estimation with the Gaze Trans former model along with MediaPipe and OpenCV for gesture recognition and face detection. A custom dataset and a subset of the MPIIGaze dataset were used for training and evaluation. Experimental results showed that initial gaze accuracy was approximately 10%, which improved significantly after preprocessing and calibration. The system achieved a mean angular error of 3.7◦, direction classification accuracy of 75.6%, and real-time responsiveness of 30 FPS. The results demonstrate that the proposed system provides an effective and intelligent hands-free document reading experience while supporting real-time human-computer interaction in practical environments.

Authors

NameRegistration No.
Namra Ghania2022-CS-550
Rehman Zahid2022-CS-514
Muhammad Ahmad2022-CS-558
Saleeha2022-CS-526
⚡

Electrical Engineering

10 Projects

Project #19 Group 1

Smart Crop Recommendation System Using Raspberry Pi 5 and Deep Learning

Supervisor Dr. Osama Bin Naeem

Abstract

Agriculture is the backbone of Pakistan’s economy, especially in Punjab, where crop productivity is heavily influenced by climatic variations. Farmers often rely on outdated knowledge, intuition, or generalized agricultural calendars, resulting in poor crop selection and reduced yield. This project proposes a Smart Crop Recommendation System that integrates real-time soil sensing, deep learning-based weather prediction, and crop recommendation modelling, designed to operate on a Raspberry Pi 5 as a low-cost, portable, and energy-efficient solution. The system uses two deep learning models. The first is a weather prediction model trained on historical climate dateset to predict climate for selected region in Punjab. The second model is a crop recommendation network that combines predicted climate data with real-time soil parameters (N, P, K, moisture, temperature, humidity) collected using integrated sensors to determine the most suitable crop. The Raspberry Pi 5 processes sensor readings, runs both models locally and displays recommendations on an LCD screen. The proposed system provides real-time, data-driven, location-specific crop guidance, enabling farmers to make informed decisions, optimize crop yield, and mitigate climate-related risks. This solution aims to contribute to sustainable agriculture, food security, and precision farming practices in Pakistan.

Authors

NameRegistration No.
Muhammad Hanzala2022-EE-608
Ammar Ghaffar2022-EE-611
Project #20 Group 2

Evaluation and Optimization of Pakistan's Transmission Network under Seasonal Load Variations

Supervisor Engr. Adnan Bashir

Abstract

This thesis presents a comprehensive operational analysis of Pakistan’s national power transmission network, focusing on congestion management, seasonal demand patterns, regional power generation distribution, and secure bulk power transfer during network emergency conditions. The study addresses the growing geographical mismatch between power generation facilities and electricity consumption centers across the country. Pakistan’s major power generation sources-coal, nuclear, thermal, and renewable energy plants-are predominantly located in the southern region, while the central and northern regions account for the majority of electricity demand. This spatial imbalance necessitates large-scale power transfer through critical transmission corridors, with Matiari Junction serving as a key hub for transmitting power from the south to the north. As a result, the corridor experiences persistent heavy loading, leading to congestion, increased transmission losses, reduced security margins, voltage instability, and vulnerability during contingency conditions. The transmission network was analyzed under multiple operational scenarios, including normal conditions, summer peak demand, winter peak demand, and extreme contingency cases. The summer scenario was identified as the most critical due to increased cooling loads, agricultural demand, and commercial activity. The results show significant increases in generation requirements, line loading, and system losses during this period, placing substantial stress on the network. A regional generation analysis was conducted by dividing the system into North, Central, and South regions. The North region relies mainly on hydropower and nuclear generation, including Tarbela and Mangla, while the Central region acts as a balancing zone with flexible thermal generation. The South region has the highest generation capacity; however, transmission constraints limit its effective utilization.Two practical solutions were proposed and evaluated. The first involves reallocating generation from the South region to the Central region, reducing transmission congestion, minimizing losses, and improving system resilience without additional capital investment. The second solution proposes the addition of a new transmission line between Matiari and Moro to enhance transfer capability, distribute loading, and improve long-term system ability. Furthermore, the study recommends the implementation of an HVDClinkbetween Matiari and Dadu to support efficient bulk power transfer and future grid expansion. The findings demonstrate that optimized generation dispatch combined with strategic transmission reinforcement can significantly enhance the reliability, efficiency, and secure operation of Pakistan’s interconnected power system.

Authors

NameRegistration No.
Awais Rafique2022-EE-607
Ali Ahmad2022-EE-613
Muhammad Abdullah2022-EE-621
Project #21 Group 3

Piezoelectric-Based Smart Helmet Accident Detection System

Supervisor Engr. Amna Javed

Abstract

Motorcycle accidents continue to claim thousands of lives annually in Pakistan and across the developing world. The critical window of survival — commonly known as the “Golden Hour” in emergency medicine — is frequently missed due to the victim’s inability to call for help, particularly in cases involving unconsciousness, severe trauma, or accidents occurring in isolated areas. While conventional helmets provide essential physical protection to the rider’s head, they lack any form of intelligent post-accident communication or automated emergency response capability. This thesis presents the design, development, and implementation of a Piezoelectric-Based Smart Helmet Accident Detection System — an embedded, IoT-enabled solution that autonomously detects motorcycle accidents and dispatches emergency alerts in real time. The system employs a dual-sensor fusion strategy combining a Piezoelectric Impact Sensor with an MPU6050 Accelerometer/Gyroscope module to achieve accurate and reliable accident detection. The piezoelectric sensor generates a measurable voltage upon detecting sudden mechanical impact on the helmet, while the MPU6050 simultaneously measures abnormal changes in acceleration, angular rate, and orientation indicative of a crash event. Upon joint threshold exceedance by both sensors, the system initiates a 5-second warning phase activated via a buzzer and vibration motor, during which the rider may press a cancel button to suppress a false alarm. If no cancellation is received within this window, the system queries the Neo-6M GPS module to obtain precise geographic coordinates (latitude and longitude) and formats them into an emergency SMS message that is transmitted via the SIM800L GSM module to a set of preconfigured emergency contacts. Additional safety features include a helmet-wearing detection mechanism using a Force Sensitive Resistor (FSR) or Reed Switch, which ensures the system is active only when the helmet is being worn, thereby preventing unnecessary false activations. Furthermore, a voice-based SOS trigger system allows the rider to vocally activate the emergency protocol by speaking predefined distress keywords such as “Help,” “Emergency,” or “SOS.” This feature is particularly valuable in scenarios where the rider is conscious but physically incapacitated. The system is powered by a rechargeable Li-ion battery managed through a TP4056 charging module, ensuring portability and long operational life. The complete system has been integrated within a standard motorcycle helmet and tested under simulated accident scenarios, yielding accurate detection, minimal false alarm rates, and reliable SMS delivery. Experimental results demonstrate 100% accident detection accuracy for severe impacts, zero false alarms across all non-accident test conditions, 95% SMS delivery success rate, and a battery life exceeding 35 hours of continuous monitoring. The proposed solution represents a cost-effective, scalable, and practically deployable advancement in personal road safety technology for the motorcycle-riding population. With a total component cost of approximately PKR 6,310 (approximately USD 22), the system is economically viable for widespread deployment in Pakistan and similar developing markets.

Authors

NameRegistration No.
Farheen Fayyaz2022-EE-604
Waseem Ahmed2022-EE-619
Raja Sanaullah2022-EE-620
Project #22 Group 4

Solar Wireless Charging & Monitering System of EV's.

Supervisor Engr. Haseeb

Abstract

The main idea behind a solar wireless EV charging system is to save time on charging using the theory of electromagnetic induction. The working of such a charging station involves transferring electrical energy collected through solar panels in a wireless fashion to the electric vehicle while the latter moves. Some of the most important parts of the charger are battery, regulation circuit, copper coils, boost converters, LED displays, and solar panels which allow the process of energy transfer to be carried out successfully. Thus, it becomes evident that charging can take place in a wireless manner without any plug-in stations at all.

Authors

NameRegistration No.
Muhammad Tasawar2022-EE-615
Ahsan Ali2022-EE-616
Ali Hamza2022-EE-617
Project #23 Group 5

Design of a Multichannel Leaky Wave Antenna for 5G Bands

Supervisor Dr. Waqas Tariq Toor

Abstract

The rapid evolution of wireless communication systems, particularly 5G and emerging terahertz (THz) technologies, has increased the demand for transmission line and antenna structures with wide bandwidth, low loss, strong field confinement, and efficient impedance matching. Conventional transmission line technologies such as Microstrip, SIW, and CPW exhibit significant limitations at high frequencies, including radiation losses, limited bandwidth and impedance mismatch, making them less suitable for modern applications. Overall, the design offers a compact, low-loss and wideband solution, making it suitable for 5G systems, microwave components, radar and THz applications, with future work aimed at experimental validation and reconfigurable designs. This work presents the design and analysis of a wideband leaky-wave antenna based on a CPWG-fed Spoof Surface Plasmon Polaritons (SSPP) structure. The SSPP mechanism enables highly confined surface wave propagation through engineered periodic structures, while a tapered CPWG-to-SSPP transition ensures efficient mode conversion with minimal reflection. Additionally, split-ring resonators (SRRs) are incorporated to achieve multiband frequency selectivity. The proposed design is simulated using HFSS on a Rogers RT/duroid 5880 substrate. Results demonstrate a stable passband with low insertion loss (S21 ≈ 0 dB) and excellent impedance matching (S11 < −20 dB). Parametric analysis confirms the tunability of the structure for multiband applications.

Authors

NameRegistration No.
Amna Bazzal2022-EE-601
Muskan Anwar2022-EE-603
Muhammad Sohail Riasat2022-EE-618
Project #24 Group 6

Solar Powered Intelligent Flood Monitoring System with Machine Learning Prediction

Supervisor Engr. Shahbaz Bashir

Abstract

Flooding remains one of the most destructive and frequently occurring natural disasters worldwide, causing significant loss of life, infrastructure damage, and economic disruption. Conventional flood monitoring systems often rely on manual observation or static threshold-based alerts, which lack the predictive capability required for timely disaster response. This project presents the design and implementation of a Solar Powered Intelligent Flood Monitoring System that integrates real-time sensor data acquisition with a machine learning-based prediction framework to enable proactive flood risk assessment. The proposed system employs an ultrasonic sensor for continuous water level measurement, alongside temperature and humidity sensors to capture critical environmental parameters that influence flood dynamics. These sensors interface with an embedded microcontroller that facilitates real-time data collection and transmission. To ensure sustainable and uninterrupted operation in remote or off-grid locations, the system is powered entirely by a solar energy harvesting unit, incorporating a photovoltaic panel coupled with a battery storage module for continuous functionality during low-light conditions. A key contribution of this work is the independently trained machine learning model, developed using historically collected real-time sensor readings. The model was trained, validated, and tested on datasets directly sourced from the deployed sensor network, ensuring domain-specific accuracy and relevance. The predictive algorithm analyzes patterns in water level fluctuations alongside atmospheric conditions to forecast potential flood events with measurable lead time, enabling early warning dissemination to relevant authorities and communities. Experimental results demonstrate that the system achieves reliable prediction performance with high accuracy, while the solar-powered architecture ensures operational continuity without dependence on conventional power infrastructure. This solution presents a cost-effective, energy-autonomous, and intelligent approach to flood risk mitigation, with practical applicability in flood-prone rural and semi-urban environments. Keywords :- Flood Monitoring, Machine Learning, Ultrasonic Sensor, IoT, Solar Energy, Early Warning System, Predictive Analytics, Environmental Sensing

Authors

NameRegistration No.
Emman Fatima2022-EE-602
Abdul Wasay2022-EE-609
Muhammad Junaid2022-EE-622
Project #25 Group 7

IoT-Based Smart Biofloc Aquaculture System for fish farming in Pakistan

Supervisor Dr. Rana Tariq

Abstract

Biofloc Aquaculture is one of the fastest-growing agricultural sectors in Pakistan and plays an important role in food production, employment, and economic development. However, traditional fish farming methods face several challenges including poor water quality management, high feed cost, disease outbreaks, excessive water consumption, and limited availability of suitable land resources. These issues significantly reduce fish productivity and increase operational costs for farmers. To overcome these limitations, Biofloc Technology (BFT) has emerged as an innovative and sustainable aquaculture technique that improves water quality through the conversion of organic waste into microbial bioflocs, which additionally serve as a natural protein-rich food source for fish. Despite its advantages, Biofloc systems require continuous monitoring and precise control of environmental parameters because even minor variations in water quality can negatively affect fish health and microbial stability. Parameters such as dissolved oxygen (DO), pH, temperature, turbidity, and ammonia concentration must remain within safe operational ranges to ensure proper fish growth and survival. Manual monitoring methods are ofteninefficient, time-consuming, and unable to provide real-time corrective action, resulting in fish stress, mortality, and financial losses. This project presents the design and implementation of an IoT-based Biofloc Aquaculture Monitoring and Control System with Real-Time Data Acquisition. The proposed system utilizes an ESP32 microcontroller integrated with multiple environmental sensors including dissolved oxygen, pH, temperature, and turbidity sensors for continuous monitoring of water quality conditions. The acquired sensor data is processed in real time and transmitted to a cloud-based IoT platform through Wi-Fi communication for remote monitoring and data logging. The developed system also incorporates automated control mechanisms through relay controlled actuators such as aerator pumps and water circulation pumps. Whenever the measured environmental parameters deviate from predefined threshold values, the system automatically activates the required corrective devices without human intervention. This intelligent monitoring and control mechanism helps maintain stable water conditions, improves fish survival rate, reduces labour dependency, and enhances overall aquaculture productivity. In addition, machine learning concepts and data analysis techniques can be integrated with the collected sensor data to predict fish health conditions, detect abnormal environmental trends, and optimize feeding and aeration strategies. The proposed system specifically focuses on Tilapia fish farming under the environmental conditions of Pakistan, where Biofloc technology has strong potential for sustainable aquaculture development. Experimental results demonstrate that the implemented system provides stable real-time monitoring, accurate sensor measurements, reliable wireless communication, and effective automated control performance. The use of Biofloc technology combined with IoT and automation improves feed conversion efficiency, reduces water consumption, minimizes fish mortality risk, and increases overall production efficiency. Therefore, this study concludes that the proposed IoT-based Biofloc monitoring and control system offers a low-cost, energy-efficient, and intelligent solution for sustainable fish farming in Pakistan. The developed system can support small-scale as well as commercial aquaculture farms by improving environmental monitoring, reducing operational costs, and increasing fish production using limited natural resources.

Authors

NameRegistration No.
Talha Zafar2022-EE-606
Abdullah Zahid2022-EE-610
Abdul Wahab2022-EE-614
Project #26 Group 8

EEW-Enabled Earthquake Safety Bed System

Supervisor Engr. Amna Javed

Abstract

Earthquakes rank among the deadliest natural disasters which occur in tectonically active areas of Pakistan through the Indian and Eurasian plate boundaries which produce continuous seismic activity. The 2005 Kashmir earthquake (Mw 7.6) recorded more than 86000 deaths because many people died during the night when residents slept and could not react to danger. The safety beds that exist for the safety of people during an earthquake depend only on local vibration detection systems that have a number of drawbacks. The system experiences multiple issues which include its need to wait until S-waves arrive at the location before it can start protecting users during the initial two seconds of earthquake vibrations and its tendency to produce false alarms from non-seismic events. Our solution proposes the first-ever safety bed system that incorporates Early Earthquake Warning (EEW) technology in Pakistan. The proposed safety bed system incorporates information obtained from P-wave detection networks, which is later validated through a high-precision accelerometer system, where a robust dual trigger system is used to automatically enclose the occupants within a protective cage through linear actuators. Key innovations include cloud-based EEW integration via the ESP32 microcontroller, MPU6050-based secondary confirmation to eliminate false positives, battery backed operation for power outages, and low cost, locally sourced components suitable for mass adoption in developing countries. The system is designed to provide advance warnings between 10 and 60 seconds which operates better than systems that rely solely on vibration detection. The research presents a complete overview of the system including its architectural design methods and component selection process and its operational procedures and expected benefits to society.

Authors

NameRegistration No.
Muhammad Rehan Majeed2021-EE-603
Rana Muhammad Ahad2021-EE-617
Muhammad Sufyan2021-EE-628
Project #27 Group 9

Autonomous Robotic Waiter With Integrated Billing Enhancing Resturant Automation And Efficiency

Supervisor Dr. Osama Bin Naeem

Abstract

An Autonomous Robotic Waiter System is designed and implemented as an end-to-end solution for handling food delivery, table detection, live detection, and automatic billing. The implemented system solves the issue of labor shortage, poor service quality and decreasing overhead costs for restaurant automation with affordable price. The Robot uses ESP8266 for controlling movements with infrared sensors for line tracking, ESP Camera module to detect the color of table for recognition based on Machine Learning, and a Thermal Printer is used to print bill on the same location where order is provided. ESP8266 Microcontroller based robot interacts with admin Dashboard and Android application through Wifi while Firebase realtime database is used for syncing realtime data amongst ordering android application, delivery robot and billing android application. Automatic Billing feature along with Live video streaming performance of robot and Quick Call-to-Action Help (Q-Help) feature contributes towards seamless end-to-end automation of restaurant with high reliability. Experimentation helps to conclude the Robot's performance towards achieving navigation and localization, efficient communication between ordering application to delivery bot and billing application alongwith detection and localization of orders. Keywords: Autonomous Robot, Restaurant Automation, Food Delivery, Bill Printing, IoT, Real-Time Monitoring, Robot Navigation, Mobile Application, Thermal Printer, Line Following, Customer Interaction, Robot Localization, Wireless Communication, Real-Time Video Streaming.

Authors

NameRegistration No.
Syed Hunain Zaidi2021-EE-604
Abdul Rehan2021-EE-629
Ayesha Shamim2021-EE-642
Project #28 Group 10

Perceptron Robotics: An AI-Powered MobileRobot for Smart Waste Sorting and Assistive Object Handling

Supervisor Dr. Osama Bin Naeem

Abstract

Public places such as shopping malls, hospitals, universities, offices, and restaurants generate a significant amount of daily waste, including bottles, cups, and other disposable objects. Manual collection and sorting of such waste is repetitive, time-consuming, and often unhygienic. In addition, elderly and physically disabled individuals may face difficulty in picking and placing objects independently. To address these challenges, this project presents Perceptron Robotics: An AI-Powered Mobile Robot for Smart Waste Sorting and Assistive Object Handling. The developed system is an AI-powered mobile robotic platform capable of detecting, approaching, picking, transporting, and sorting selected objects into assigned bins. The robot uses an external wired camera mounted at the front side above the ultrasonic sensor to capture live video. The video stream is processed using YOLOv8 and OpenCV for real-time object detection and classification. The target objects considered in this project are bottle, cup, and pink object. Each object is assigned to a specific colored bin: bottle to the red bin, cup to the yellow bin, and pink object to the green bin. The system uses ROI-based navigation to guide the robot toward the detected object. If the object is not centered within the region of interest, directional commands are generated and sent to the hardware system through Firebase Realtime Database. Once the object reaches the required position, the robot moves toward it and stops using an ultrasonic sensor when the object is detected in front of the robot. The robotic arm then moves downward, opens the claw, grips the object, and lifts it. After picking the object, the system tracks the relevant colored bin using HSV-based color detection and navigates the robot toward the assigned bin. Finally, the robotic arm drops the object into the correct bin. The system can operate in two processing modes: laptop-based processing through Py Charm and embedded processing using Raspberry Pi 4. The hardware implementation includes NodeMCU, Arduino Nano, DC motors, motor drivers, ultrasonic sensor, servo motors, PCA9685 PWM servo driver, robotic arm, gripper, and Firebase-based wireless communication. Experimental testing was performed through 10 trials, out of which 9 trials were successful, resulting in an overall success rate of 90%. The main failure causes observed during testing were gripper misalignment, lighting variations, and network/WiFi vii communication delay. The proposed system demonstrates a low-cost practical prototype for smart waste sorting and assistive object handling. It combines artificial intelligence, computer vision, embedded systems, wireless communication, and robotic manipulation to provide an automated solution for public waste management and object handling assistance.

Authors

NameRegistration No.
Khalil Ahmed2021-EE-624
Iqra Arshad2021-EE-649
🏗️

Civil Engineering

5 Projects

Project #29 Group 1

Soil Enhancement Techniques for Foundation in Diverse Geographical Regions of Pakistan

Supervisor Dr. Khawaja Adeel Tariq

Abstract

Foundation failures are frequently observed in regions such as Narowal, Kashmir (Satellite Town Thotha, Muzaffarabad), and Gilgit Baltistan (District Ghanche, Village Keris) due to weak and variable subsoil conditions. Structural issues including wall cracking and differential settlement are common in these areas because the underlying soils are unable to adequately sustain structural loads. This study evaluates the engineering characteristics of soils collected from the selected regions and investigates economical and sustainable soil enhancement techniques for improving foundation stability. Laboratory investigations included sieve analysis, Atterberg limits, soil classification, Modified Proctor compaction, and CBR tests. The test results indicated that the majority of the soil samples belonged to the SM, SP, SP-SM, and SC categories and were generally poorly graded with varying fines content, leading to low bearing capacity, moisture sensitivity, and settlement-related problems. To improve the engineering performance of the soils, stabilization was carried out using locally available waste and by-product materials including fly ash, rice husk ash, lime, and cement. Industrial visits were also conducted to identify the availability and applicability of such by-products within the selected regions. Experimental findings showed that fly ash produced the most effective improvement in soil behavior by enhancing particle gradation, increasing dry density, and improving strength characteristics. Rice husk ash reduced the optimum moisture content; however, its strength improvement was comparatively limited when used independently. Lime effectively reduced plasticity and moisture sensitivity, whereas cement was utilized only for comparative laboratory evaluation and was not considered a preferred stabilizer due to its relatively high cost. In comparison with expensive foundation solutions such as raft and deep foundation systems, the proposed stabilization techniques provide a more economical and environmentally sustainable alternative through the utilization of industrial and agricultural waste materials. The findings of this research provide practical guidance for engineers and practitioners in selecting region-specific soil improvement methods for safe, durable, and sustainable foundation construction practices in Pakistan.

Authors

NameRegistration No.
Ayesha Latif2022-CIV-301
Usama Dilpazeer2022-CIV-303
Noman Jaffar2022-CIV-304
Abideen Ali2022-CIV-305
Project #30 Group 2

Development of cementless concrete mixture for casting of full scale spun cast concrete pipe

Supervisor Dr. Adeel Faisal

Abstract

This project was carried out to develop a cementless geopolymer concrete (GPC) mixture for casting full-scale spun cast concrete pipes as a sustainable alternative to conventional cement concrete. The main idea behind the study was to reduce the use of Ordinary Portland Cement because of its environmental impacts and replace it with fly ash-based geopolymer concrete. In this research, fly ash was used as the source material, while sodium hydroxide (NaOH) and sodium silicate (Na₂SiO₃) were used as alkaline activators. The work was completed in two phases. In the first phase, the NaOH concentration was kept 8M and the target slump was maintained around 25 mm. Different laboratory specimens were prepared including 8 PCC cylinders, 8 GPC cylinders, 2 prisms, and 12 fiber-reinforced GPC cylinders. Among the fiber-reinforced samples, four cylinders contained 1% steel fibers, four contained 1% polypropylene (PP) fibers, while the remaining four had a hybrid combination of 0.5% steel fibers and 0.5% PP fibers. These specimens were cast for compressive and flexural strength testing at 28 and 56 days. However, the results obtained from the first phase were not satisfactory. The concrete did not gain the required strength properly, mainly because excessive heat was generated during the preparation of the chemical solutions, which affected the geo-polymerization process. After analyzing the first phase results, modifications were made in the mix design during the second phase. The concentration of NaOH solution was increased from 8M to 12M, and both the activator-to-fly ash ratio and NaOH-to-Na₂SiO₃ ratio were also increased. In this phase, four full scale spun cast concrete pipes were cast having a length of 8 ft, internal diameter of 1 ft, and wall thickness of 2 in. Pipe 1 was reinforced with a conventional steel cage, while Pipes 2, 3, and 4 were reinforced using fibers only without any steel cage. These included 1% steel fibers, 1% PP fibers, and a hybrid mix of 0.5% steel fibers with 0.5% PP fibers, respectively. Along with the pipes, companion cylinders and prisms were also cast. The testing of the second phase is still in progress and the final results are yet to be evaluated.

Authors

NameRegistration No.
M. Muzammil Mustfa2022-CIV-311
Shah Fahad2022-CIV-313
M. Rizwan2022-CIV-321
Saqib Sheikh2022-R/2021-Civ-340
Project #31 Group 3

Characterization of soil properties and bearing capacity evaluation for urban & rural zones of narowal district Pakistan

Supervisor Dr. Khawaja Adeel Tariq

Abstract

Narowal District, located in the Punjab province of Pakistan, is experiencing rapid urbanization, population growth, and increasing infrastructure development. However, the region lacks updated and reliable geotechnical information regarding soil properties and bearing capacity, which are essential for safe and economical foundation design. The existing soil maps and available data are outdated and insufficient for supporting modern construction practices, land-use planning, and sustainable development. Inadequate assessment of soil conditions can lead to excessive settlement, structural instability, foundation failure, and increased construction and maintenance costs. Therefore, this study aims to characterize the soil properties and evaluate the bearing capacity of selected urban and rural zones of Narowal District. To achieve these objectives, soil samples were collected from more than fifty representative locations distributed across urban, rural, and boundary areas of the district. The sampling depth was maintained at approximately 3 feet, representing typical shallow foundation conditions commonly used in residential and low-rise structures. Laboratory testing was carried out to determine the physical and engineering properties of the soil. The tests conducted included sieve analysis, Atterberg limits, Modified Proctor Test, consolidation test, and direct shear test. These tests were used to evaluate particle size distribution, plasticity characteristics, compaction behavior, compressibility, and shear strength parameters of the soil. Based on the obtained results, soils were classified using standard geotechnical classification systems, while GIS tools were utilized to develop a surface soil zoning map illustrating the spatial distribution of soil properties across the district. The findings of this study provide a localized geotechnical database and preliminary bearing capacity profiles for Narowal District. The results will assist engineers, planners, architects, and developers in selecting suitable foundation systems and making informed construction decisions. Furthermore, this research contributes to safer, more economical, and sustainable infrastructure development and can serve as a reference for future geotechnical investigations and urban planning projects in similar regions of Pakistan.

Authors

NameRegistration No.
Kashif Ali2022-CIV-306
Waheed Iqbal2022-CIV-309
Hussnain Ali2022-CIV-314
Saif Ameer2022-CIV-318
Project #32 Group 4

Mesoscale evaluation of bond behavior and fracture characteristics of concrete employing coupled RBSM and solid FEM model

Supervisor Dr. Shoaib Karam

Abstract

The bond behavior between deformed steel reinforcement and concrete is a critical factor governing the structural performance of reinforced concrete (RC) members. While numerous experimental investigations have been conducted to characterize this bond behavior at the surface level, the detailed internal fracture mechanisms of concrete under axial loading remain insufficiently understood. This study presents a mesoscale numerical investigation of bond behavior and fracture characteristics of concrete using a coupled Rigid Body Spring Model (RBSM) and solid Finite Element Method (FEM) framework. In the proposed numerical model, concrete is discretized using a three-dimensional RBSM based on the formulation of Kawai (1977), incorporating constitutive models developed by Yamamoto et al. (2008) to capture nonlinear mechanical responses including crack initiation, propagation, and failure. Deformed steel rebars of diameters D19 mm and D32 mm are modeled using eight noded solid FEM elements, with actual geometrical details of the rebar ribs explicitly accounted for. The two models are coupled at the steel-concrete interface through link elements consisting of normal and shear springs, enabling an accurate representation of interfacial stress transfer. The study investigates the influence of key parameters including rebar diameter, lug spacing, and lug geometry on bond strength, internal crack formation, and failure modes. The numerical model is validated against existing experimental data through comparison of crack patterns and fracture progression under varying steel stress levels. The results demonstrate that the coupled RBSM FEM model effectively captures internal cracking behavior, secondary crack development, and the characteristic comb-like fracture structures observed experimentally. This research provides quantitative insight into the internal fracture mechanism of RC specimens, contributing to improved design guidelines for deformed reinforcement in concrete structures.

Authors

NameRegistration No.
Saeed Ahmad2022-CIV-312
Qadeer Maqbool Mughal2022-CIV-316
M. Khubaib Ishaq2022-CIV-317
Moeez Haidar2022-CIV-320
Project #33 Group 5

Impact of climate change on runoff generation of indus basin

Supervisor Engr. Sehrish Khan

Abstract

Climate change is increasingly affecting runoff generation within the Upper Indus Basin through rising temperatures, glacier retreat, and changing precipitation patterns. Variations in snow accumulation and snowmelt timing are altering seasonal river discharge behaviour, directly influencing hydropower generation and water-resource management in Pakistan. This study evaluates the impacts of climate change on runoff generation within the Upper Indus Basin catchment using the Hydrological Predictions for the Environment (HYPE) model. Catchment delineation and terrain analysis were carried out using HydroSHEDS, HydroBASINS, and USGS SRTM1 Digital Elevation Model (DEM) datasets within ArcGIS Pro and QGIS environments. The modelled catchment area covered approximately 66,830 km² within the accessible region of Pakistan. Due to geopolitical and data-access limitations, portions of the Upper Indus Basin extending into Jammu & Kashmir, India, and China were excluded from the analysis. Land-cover classification and GeoClass preparation were performed using ESA WorldCover datasets, while precipitation and temperature data were obtained from CMIP6 climate projections under the SSP2-4.5 scenario. Climate forcing datasets including Pobs and Tobs files were generated using Python libraries such as xarray and pandas. Preliminary simulations indicated seasonal variability in runoff generation, with increased summer discharge associated with accelerated snow and glacier melt. Simulated mean discharge values ranged approximately between 750–1000 m³/s, while observed regional discharge trends generally varied between 700–1100 m³/s. The difference between simulated and observed discharge values remained within an estimated range of 5–15%, indicating acceptable preliminary agreement for a prototype-scale HYPE implementation. The study provides a scientific framework for climate-resilient hydrological assessment and sustainable water-resource planning in Pakistan.

Authors

NameRegistration No.
Dua Babar2022-CIV-307
Rao M. Qasim2022-CIV-315
M. Tayyab Farooqi2022-CIV-319
Rohan Babar2023-R/2022-CIV-310
⚙️

Mechanical Engineering

8 Projects

Project #34 Group 1

Fabrication and implementation of Sensor based HE-OBCU-EGR emission Control Unit for CI Engines

Supervisor Engr. Umar Ishaq

Abstract

The continuous rise in vehicular pollution and stringent environmental regulations have created an urgent need for efficient and sustainable exhaust emission control technologies. This project presents an advanced exhaust emission control unit designed to reduce harmful emissions and improve air quality through effective exhaust gas purification and recirculation. The system is capable of lowering exhaust gas temperature, removing soot particles, trapping decomposition products, and supplying cleaner exhaust gases back to the engine through an Exhaust Gas Recirculation (EGR) process. By reducing toxic emissions such as nitrogen oxides (NOx), unburnt hydrocarbons, and particulate matter, the unit contributes to cleaner combustion and environmentally friendly engine operation. The proposed technology provides a compact, economical, and efficient solution for modern emission control requirements while supporting the development of sustainable transportation systems. In addition to improving engine performance and reducing atmospheric pollution, the system promotes energy-efficient utilization of exhaust gases, thereby helping to create a greener and healthier environment. Future developments of the proposed unit may include the use of smart sensors for automatic emission monitoring, improved filtering oils for better contaminant absorption, and lightweight materials to enhance durability and efficiency. These advancements can further increase emission reduction capability and support the global transition toward cleaner automotive technologies and an emission-free climate.

Authors

NameRegistration No.
Afraz2022-ME-518
Ahtisham2022-ME-508
Saim2022-ME-513
Sania Kanwal2022-ME-501
Project #35 Group 2

Fabrication of an Intelligent Hybrid Thermal Management System for EV Batteries using CFD Analysis and Machine Learning-Based Adaptive Cooling

Supervisor Dr Zahid Hussain

Abstract

Lithium-ion batteries are widely used in electric vehicles, portable electronic devices, and renewable energy storage systems because of their high energy density and long cycle life. However, excessive heat generation during charging and discharging processes negatively affects battery performance, thermal stability, safety, and lifespan. Conventional cooling techniques such as air cooling and liquid cooling suffer from limitations including poor thermal uniformity, hotspot formation, and high energy consumption under high thermal loads. To overcome these limitations, the present study proposes an intelligent hybrid battery thermal management system integrating Computational Fluid Dynamics (CFD), machine learning (ML), and prototype implementation for efficient thermal regulation and energy optimization. The proposed system combines liquid cooling and air cooling using a dual water-body cooling configuration consisting of cooling channels, coolant circulation loop, radiator, cooling fan, and adaptive thermal control mechanism. CFD simulations were performed using ANSYS Fluent 2022 R1 to analyze temperature distribution, velocity behavior, pressure characteristics, and heat transfer performance within the cooling domain. The numerical results demonstrated effective thermal management with maximum battery temperature maintained near 300 K and minimum temperature approaching 298 K, indicating improved thermal uniformity and reduced hotspot formation. Velocity analysis showed stable coolant circulation within the range of 2.0–3.45 m/s, while pressure contours confirmed smooth coolant flow behavior. An Artificial Neural Network (ANN)-based machine learning model was developed using CFD-generated datasets to predict battery thermal behavior and provide intelligent fan and pump control according to thermal demand. The adaptive cooling strategy achieved approximately 40% reduction in energy consumption compared with conventional cooling systems. Prototype validation confirmed good agreement with CFD results, demonstrating the effectiveness of the proposed intelligent hybrid cooling framework for advanced lithium-ion battery thermal management applications.

Authors

NameRegistration No.
Shehryar Altaf2022-ME-511
Hasnain2022-ME-507
Burhan2022-ME-525
Hamna2022-ME-502
Project #36 Group 3

Fabrication and Implementation of Smart Safety System (SSS) for traditional Chaff Cutter Model using sensor fusion and Automatic Braking Control Mechanism

Supervisor Engr. Basit Ali Wajid

Abstract

This project focuses on enhancing the safety of traditional flywheel type chaff cutters widely used in rural agricultural environments. Although these machines are efficient and affordable, they pose serious risks due to their open feeding mechanism and high-speed rotating blades. Accidents often occur when operators’ hands come too close to the cutting zone, and the situation is worsened by the delayed stopping time caused by flywheel inertia. To address these challenges, a Smart Safety System (SSS) is proposed that integrates vision based hand detection with an automatic disk braking mechanism. A camera is installed near the feeding chute to monitor the operator’s hand movement in real time using image processing techniques such as skin color segmentation and contour detection. Additionally, a color sensor is used as a secondary validation system to improve detection accuracy under varying lighting conditions. When a potential hand intrusion is detected within a predefined danger zone, the system immediately activates an electromechanical disk brake mounted on the flywheel shaft. This braking mechanism rapidly reduces the rotational speed and brings the system to a stop within a fraction of a second, significantly minimizing injury risk. The system also includes additional safety features such as a blade cover and an emergency stop button to provide multiple layers of protection. Designed as a costeffective retrofit solution, the system can be easily installed on existing machines without major modifications. Overall, this project offers a practical and innovative approach to improving operator safety in rural agricultural settings while maintaining affordability and efficiency.

Authors

NameRegistration No.
Kashif2022-ME-504
Hamza2022-ME-516
Islam2022-ME-521
Project #37 Group 4

CFD-Based Optimization and Hydrodynamic Analysis of an Efficient Wave Energy Converter

Supervisor Engr. Muhammad Lolak

Abstract

This study presents a CFD-based optimization and hydrodynamic analysis of a Wave Energy Converter (WEC) aimed at improving wave energy extraction efficiency through buoy geometry optimization. Ocean wave energy is considered a promising renewable energy source due to its high energy density and sustainability. However, the efficiency of many existing WEC systems is limited because of non-optimized buoy designs that reduce effective interaction with waves. The project focuses on investigating the influence of buoy geometry on hydrodynamic performance using Computational Fluid Dynamics (CFD) simulations. Three different buoy configurations were designed in SolidWorks, including a classical bulbous-bottom buoy, an axisymmetric top-shaped buoy, and an innovative multi-diameter buoy. Each geometry was analyzed under identical wave conditions in a numerical wave tank using a two-phase air–water model with the Volume of Fluid (VOF) method. The simulations evaluated parameters such as stability, buoyancy, excitation force, and wave interaction characteristics. Comparative hydrodynamic analysis showed that buoy geometry significantly affects system performance. Buoy 1 demonstrated balanced stability with moderate energy interaction, while Buoy 2 showed high responsiveness but poor stability due to negative metacentric height. Buoy 3 achieved the best overall performance, exhibiting strong wave interaction, higher excitation force, and stable hydrodynamic behavior. Based on the CFD results, Buoy 3 was selected as the optimal configuration for maximum energy extraction potential. The project also includes the development of a physical prototype for future experimental validation. The findings confirm that CFD-based geometric optimization can significantly enhance the efficiency and stability of wave energy converters, contributing to the advancement of sustainable marine renewable energy technologies.

Authors

NameRegistration No.
M Haseeb2022-ME-512
M Asad2022-ME-520
M Haider2022-ME-524
Project #38 Group 5

Computational and Experimental Analysis for Optimizing Molten Steel Refining in Ladle Furnace to Enhance Industrial Efficiency

Supervisor Dr Zahid Hussain

Abstract

In secondary steelmaking, efficient gas stirring inside the ladle furnace plays an important role in improving steel quality, refining efficiency, and process stability. In this study, a combined computational and experimental approach was used to investigate the hydrodynamic behavior of a gas-stirred ladle system. The work mainly focuses on plume development, circulation patterns, turbulence generation, slag eye formation, slag layer thickness, and mass transfer behavior under different gas stirring conditions. Computational Fluid Dynamics (CFD) simulations were performed using both Euler–Euler and Euler–Lagrange approaches to study the interaction between argon bubbles, molten steel, and slag layer inside the ladle. To validate the numerical results, a 1:4 downscaled water model was developed based on geometric, kinematic, and dynamic similarity criteria. The experimental setup consisted of an acrylic vessel, air injection system, velocity probes, tracer dye, and camera-based optical flow analysis using Python and OpenCV. Experimental observations were compared with CFD predictions to evaluate plume behavior, velocity distribution, turbulence intensity, and mixing characteristics. Good agreement was observed between the simulation and experimental results. The study showed that gas flow rate, bubble size, and slag layer thickness strongly influence the overall refining performance. Low gas flow rates resulted in weak circulation and poor mixing, while very high flow rates created excessive turbulence, unstable slag eye formation, and possible slag entrainment near the free surface. Moderate gas flow rates produced stable circulation patterns and improved mixing efficiency. It was also observed that thicker slag layers helped maintain interfacial stability and controlled excessive slag eye expansion, leading to better mass transfer conditions. Among the two numerical approaches, the Euler–Lagrange model provided more realistic predictions of bubble motion and turbulent flow behavior. Overall, this work provides a reliable framework for optimizing gas-stirring practices in ladle metallurgy and improving industrial steel refining operations.

Authors

NameRegistration No.
M Adnan2022-ME-522
Waqas Nasir2022-ME-515
Saira Azam2022-ME-528
Project #39 Group 6

IoT Enabled Performance Optimization of Helical Blade Vertical Axis Wind Turbine

Supervisor Engr. Asif Jalal

Abstract

The increasing demand for sustainable and clean energy has accelerated the development of renewable energy technologies, particularly wind energy systems. This project presents the design, analysis, and implementation of an IoT-enabled performance optimization system for a helical blade Vertical Axis Wind Turbine (VAWT). The proposed turbine utilizes a three-blade helical configuration based on the NACA 0018 airfoil profile to achieve smooth rotational motion, reduced torque fluctuation, and improved aerodynamic efficiency under low and variable wind conditions. Computational Fluid Dynamics (CFD) simulations and structural analyses were performed using ANSYS and SolidWorks to evaluate airflow behavior, pressure distribution, torque generation, stress distribution, and structural reliability of the turbine. Experimental analysis was conducted at different wind velocities to study the effects of blade length and rotor radius on turbine performance. An IoT-based monitoring system was integrated into the turbine using Arduino Uno and NodeMCU ESP8266 microcontrollers along with multiple sensors, including an anemometer, Hall effect sensor, INA219 voltage-current sensor, LM35 temperature sensor, and vibration sensor. These sensors continuously monitor critical parameters such as wind speed, rotational speed (RPM), voltage, current, temperature, and vibration levels. The collected data are transmitted wirelessly to the ThingSpeak cloud platform for real-time visualization, remote monitoring, and historical data analysis. The developed monitoring system enables predictive maintenance, fault alarming and enhanced operational safety through continuous supervision and automated protection mechanisms. The results demonstrate that combining aerodynamic optimization with IoT-based smart monitoring significantly improves the efficiency, reliability, and stability of helical vertical axis wind turbines. The proposed system offers a low-cost, scalable, and intelligent renewable energy solution suitable for urban and low-wind-speed environments while contributing to sustainable energy development and Industry 4.0 applications.

Authors

NameRegistration No.
Hammad2022-ME-517
Zaman2022-ME-514
Afaq2022-ME-519
Rahat2022-ME-503
Project #40 Group 7

AI Based Condition monitoring using multiple sensors for Predictive Maintenance

Supervisor Engr. Sufyan Matloob

Abstract

This project presents the development of an AI-based condition monitoring system for rotary drive assemblies, aimed at enabling predictive maintenance and early fault diagnosis. The system integrates two key approaches: computer vision using deep learning, and vibration signal analysis using Python. In the computer vision module, a YOLOv11 object detection model was trained to automatically identify mechanical faults such as cracks in bearings and shafts, as well as misalignment issues. Annotated datasets were prepared and used to train and evaluate the model, which demonstrated high accuracy in fault localization and classification. The second module involved vibration-based analysis using time and frequency domain signals collected under various operating conditions and fault scenarios. A Python-based program was developed to visualize these signals, detect anomalies, and predict system health. Further, signal optimization techniques such as Butterworth, Chebyshev, Gaussian, and Savitzky-Golay filters were applied to enhance fault visibility. The integration of these two techniques provides a robust and intelligent framework for condition monitoring in rotating machinery. The system supports early fault detection, reduces the risk of unexpected breakdowns, and contributes to improved maintenance planning and operational efficiency.

Authors

NameRegistration No.
Tehzib Ul Hassan2022-ME-523
Umar Maqbool2022-ME-506
Umar Shahid2022-ME-510
Project #41 Group 8

Development and Fabrication of Smart Portable Solar Powered Water Filtration System using Reverse Osmosis

Supervisor Engr. Tanveer Mukhtar

Abstract

The unavailability of safe drinking water in rural and remote regions of Pakistan represents a significant public health and humanitarian challenge. Conventional water treatment systems are largely ineffective in such areas due to the absence of stable electrical infrastructure and high installation costs. This project addresses these limitations through the design, fabrication, and performance evaluation of a smart portable solar-powered water filtration system integrating Reverse Osmosis (RO) technology with renewable energy. The proposed system incorporates a sequential multi-stage purification mechanism consisting of a sediment filter, Granular Activated Carbon (GAC) filter, Carbon Block (CTO) filter, Thin Film Composite (TFC) RO membrane, and a post-treatment polishing filter. This configuration ensures comprehensive elimination of suspended particulates, dissolved salts, heavy metals, and pathogenic microorganisms. A 12V DC pump maintains the required operating pressure of approximately 9 bar across the RO membrane, powered by a 60W photovoltaic solar panel supported by dual 12V battery storage, ensuring uninterrupted off-grid operation. Structural integrity of the portable frame was validated through ANSYS-based finite element simulation, confirming deformation and stress values within permissible limits. Experimental results demonstrated substantial reductions in Total Dissolved Solids (TDS), Electrical Conductivity (EC), and pH deviation, verifying high purification efficiency. With a total fabrication cost of PKR 38,200, the system offers an economically viable, environmentally sustainable, and community-deployable solution, directly contributing to UN Sustainable Development Goals SDG 3, SDG 6, SDG 7, SDG 9, and SDG 13.

Authors

NameRegistration No.
Saqib2022-ME-509
Haseeb Tariq2022-ME-505
Fahad2022-ME-526
🏛️

Architecture

14 Projects

Project #42 Group 1

Reinterpreting Mosque Architecture for Women’s Active Participation

Supervisor Ar.Amna Iqbal

Abstract

This project explores a contemporary interpretation of mosque architecture that emphasizes inclusivity, accessibility, and women’s active engagement within sacred spaces. The design challenges conventional spatial hierarchies by creating an environment where spirituality, community interaction, education, and social connectivity coexist seamlessly.The proposal envisions the mosque as a dynamic civic and cultural institution rather than solely a place of worship. Through modern architectural expression, human-centered design, and a strong connection between faith and everyday life, the project aims to redefine the role of women in Islamic spaces while fostering a progressive and welcoming religious environment.

Authors

NameRegistration No.
Maryam Abbas2023-Arch-102
Project #43 Group 2

Interfaith House of God

Supervisor Ar.Amna Iqbal

Abstract

The Interfaith House of God is a community-centered spiritual project designed to promote harmony, coexistence, and mutual respect among different religions. The project introduces a contemporary interpretation of sacred architecture by integrating a mosque, church, and gurdwara within a single unified environment. This concept is particularly significant within the regional context, where interfaith architectural spaces remain rare.The inclusion of these three religious spaces was based on the cultural and social context of the site. The mosque represents the dominant Muslim population, while the church acknowledges the presence of Christian minority communities. The gurdwara was introduced in response to the site’s proximity to the Kartarpur Corridor, where Sikh pilgrims from around the world frequently visit. The project aims to provide these visitors with a welcoming spiritual and community space that encourages interaction and understanding beyond religious boundaries.The inclusion of these three religious spaces was based on the cultural and social context of the site. The mosque represents the dominant Muslim population, while the church acknowledges the presence of Christian minority communities. The gurdwara was introduced in response to the site’s proximity to the Kartarpur Corridor, where Sikh pilgrims from around the world frequently visit. The project aims to provide these visitors with a welcoming spiritual and community space that encourages interaction and understanding beyond religious boundaries.The design concept was inspired by water ripples, symbolizing unity, equality, and interconnectedness. Just as ripples overlap and expand seamlessly, the architectural forms and spatial arrangements were intentionally designed with overlapping geometries to express interfaith connection and coexistence. Water, light, and open transitional spaces were integrated throughout the design to enhance the spiritual atmosphere and create a sense of peace and reflection.The project also includes communal gathering spaces, public courtyards, and tourist-oriented areas that support cultural exchange and social engagement. Overall, the Interfaith House of God functions not only as a sacred space, but also as a symbol of inclusivity, dialogue, and collective identity.

Authors

NameRegistration No.
Laraib Tariq2023-Arch-101
Project #44 Group 3

Light Weight Church

Supervisor Ar.Amna Iqbal

Abstract

The project began with an initial research study on the history and evolution of Christian architecture, particularly focusing on the Anglican Church and its relationship with the traditions of the Roman Catholic Church and Protestantism. The study explored liturgical practices, spatial hierarchy, sacred symbolism, and the significance of the east–west axis in church planning. Case studies of traditional Gothic churches and contemporary sacred architecture were analyzed to understand the relationship between spirituality, form, light, and human experience. The conceptual phase focused on developing a modern interpretation of Anglican church architecture through the integration of natural light, verticality, and simplified sacred spaces. The design intent was to create a spiritually engaging environment while maintaining a contemporary architectural language inspired by selected Gothic principles. Spatial zoning and circulation were organized according to liturgical requirements, including the nave, chancel, altar, community spaces, and supporting functions. The final design evolved through conceptual sketches, zoning diagrams, and spatial development, resulting in a contemporary worship space that balances tradition and modernity. The project emphasizes light as a spiritual element, procession as a spatial experience, and architectural simplicity as a medium to enhance contemplation and communal worship. ABOUT DESIGNING The Lightwave Church is envisioned as a contemporary sacred space that brings together spirituality, community, and modern architectural expression through the interplay of form, light, and movement. Inspired by the protective geometry of shell structures and the flowing rhythm of waves, the building form symbolizes shelter, unity, and spiritual elevation. The layered curved roof creates a dynamic architectural identity while allowing natural light to filter gently into the interior spaces, reinforcing the idea of divine illumination and emotional serenity. The planning of the church is strongly based on functionality and spatial hierarchy, organizing worship, community gathering, and supportive spaces into a cohesive and efficient environment. The design focuses not only on religious practices but also on creating a welcoming community-centered atmosphere where people can gather, reflect, interact, and connect beyond formal worship activities. The rhythmic repetition of vertical façade elements and flowing roof forms generates a sense of movement and continuity throughout the project, giving the building a modern yet spiritually grounded character. The surrounding landscape is designed as an extension of the sacred environment, blending open green spaces, circulation paths, and communal areas to enhance the overall experience of calmness and reflection. Overall, The Lightwave Church represents a balance between contemporary architectural language and timeless spiritual values, creating a meaningful place where light, form, nature, and community coexist in harmony.

Authors

NameRegistration No.
Sadeed Ahmed Kamran2023-Arch-106
Project #45 Group 4

Tertiary Care Hospital

Supervisor Ar.Ahmed Iqbal

Abstract

The proposed project is an 8-acre general hospital located along Zafarwal Road, positioned as a transitional zone between rural and urban communities. The design responds directly to the social and physical needs of the surrounding population, where the majority of residents are farmers and laborers exposed to frequent physical strain and work-related injuries.While the hospital functions as a complete healthcare facility with essential medical departments and patient-care services, the core specialty of the project is a dedicated Physiotherapy and Rehabilitation Department. The design introduces advanced recovery spaces including hydrotherapy, physical exercise therapy, and rehabilitation facilities, addressing the lack of specialized treatment centers for musculoskeletal injuries within the region.The hospital was planned as a seven to eight-floor structure focused on creating a healing and patient-centered environment. Natural light, ventilation, and visual comfort were integrated throughout the design using strategically placed windows and open transitional spaces. IPD rooms were designed with separate windows and private balconies to provide patients with a calm, peaceful, and restorative atmosphere that enhances emotional and physical recovery.Spatial planning emphasizes accessibility, efficient circulation, and comfort for both patients and staff. The project ultimately aims to bridge the gap between rural healthcare limitations and modern medical infrastructure by creating a functional, therapeutic, and community-oriented healing environment.

Authors

NameRegistration No.
Laraib Tariq2023-Arch-101
Project #46 Group 5

Tertiary Care Hospital

Supervisor Ar.Ahmed Iqbal

Abstract

This project envisions a 300–320 bed tertiary care hospital that goes beyond being a medical facility, aiming instead to create a place where healing feels natural, calm, and humancentered. The design carefully organizes emergency, critical care, diagnostic, and patient spaces so that movement is clear, efficient, and stress-free for both patients and staff. Set within a 20-acre site with existing trees and open landscapes, the hospital takes advantage of nature to create a quieter, more comforting environment for recovery. The design approach is guided by ideas of simplicity, natural flow, and sensory comfort, inspired by Zen, Tai Chi, Wabi-Sabi, and phenomenological thinking. Light, shadow, texture, and material honesty are used to shape spaces that feel less institutional and more restorative. Located in Narowal, a growing district with limited access to advanced tertiary healthcare, the project responds directly to a critical regional gap in medical infrastructure. It aims to serve not only the city but also surrounding rural settlements and cross-border populations, improving healthcare accessibility in an underserved region. The presence of a large green and open site further strengthens its potential as a healing landscape, integrating nature into the recovery process. With a planned built-up area of around 32,400 sqm, the project focuses on long-term efficiency and sustainability through passive design strategies like natural ventilation, daylight use, and shaded outdoor areas. Overall, the hospital is imagined as a healing environment where advanced healthcare and human comfort work together seamlessly.

Authors

NameRegistration No.
Shahzaib Ali2023-Arch-105
Project #47 Group 6

Dwelling

Supervisor Ar.Ahmed Iqbal and Aneeka Khan

Abstract

.The project is about designing the dwelling for chosen character. In this project the dwelling must reflect the characteristics of the character. The character I chose for this project is the “Prostitute”. Prostitution is the last choice of every human being to earn. Prostitutes don’t only sell themselves but they also suffer the unimaginable pain daily.The story of the prostitute is “At the age of 20, she was kidnapped by a group of local kidnappers. She was sold to the brothel after some days. at the brothel she spends her early days in the room, she was not allowed to visit outside. whole day she stares at the window in her room where she was kept. She thinks about the freedom she holds before all this happening. she saw birds flying freely, people living life that she was living before. All this continues till six months after she was brought into business where she resists or kill herself daily to do such things daily. she saw the side of the world that no common people can ever saw. she saw the most reputative men of the society in front of her helpless by their desires and lust.”After the story the key points of her life are researched. The main key points of her life is:  Mental health.  Stucked.  Resistance.  Changed p.o.v  Freedom.  Hypocrisy.  Acceptance. Then the abstract sketches of key points are drawn to understand the basic idea and starting the thinking process. The abstract sketches provides the vague idea about the design language and spaces. After sketches abstract models are created for each key words for understanding the spaces and visualizing basic ideas for how spaces will look. Models helps to understand the relationship b/w space and light. Further the sections of the models are drawn for understanding the models in detail. Conceptual sections & plans helps to develop the design language and gives the detail of light interaction with space. Further a design language is created by the help of conceptual sketches. Then the program of the dwelling is created, that how it’s all going to work. Every space is given the specific area to perform specific function. Levels are added to make it more efficient. Then the final plans of the spces are drawn and modeled.

Authors

NameRegistration No.
Hammad Khalil2024-Arch-110
Project #48 Group 7

Voyager

Supervisor Ar.Ahmed Iqbal and Aneeka Khan

Abstract

Architectural Statement Miyamoto Musashi This Dwelling Is Designed As A Spatial Interpretation Of The Philosophy And Ronin-Saint Life Of Miyamoto Musashi . The Architecture Translates His Personality — The Starting Rash Personality And Deep Internal Solitude — Into A Sequence Of Spaces That Explore Performance, Discipline, Isolation, And Self-Realization. Concept Is Based On The Life Of Sword Saint.Architectural Narative The Story Of A Child Who Is Said To Be Possessed By Wild Beasts And Set To Achieve Mastery Of Sword. But During The Journey He Came To Know About Life , The Way , Journey , Sword, Mind . He Studied The World , The Residents And Came To Know About The Importance Of Life , But Between This Journey He Enlightened His Own Self, His Soul Path Of Self Mastery: During This Project I Came To Know About The Path Of Self Mastery Which Musashi Stated As “Practicing 1000 Days Is Discipline But Practicing 10000 Days Is Mastery And One Should See The World Through Mastery “ . His Teaching Taught Me How To Discover Your Way , Your Purpose And You Should Reach Everything With Empty Mind And Calmness He Also Said Accept Just The Way It Is Meaning What Happened And What Will Happen Is Bound To Happen One Should Not Regret About It Keyword: • Separation • Regret • Discipline • Strategy • Spirituality • Philosophy • Desire • Strength • Honour The Whole Journey Is Based On These Keywords That Covers Miyamoto Musashi’s Life And His Journey From A Masterless Samurai(Ronin) To Sword Saint

Authors

NameRegistration No.
Hafiz subhan2024-Arch-124
Project #49 Group 8

Journey through humanity

Supervisor Ar.Ahmed Iqbal and Aneeka Khan

Abstract

The design process began with abstract sketches where random lines, fragmented geometries, textures, and free compositions were explored to study movement, emotion, and spatial interaction. The aim of these initial explorations was not simply to create shapes, but to discover hidden spatial possibilities within abstract forms. Through continuous experimentation, these sketches gradually revealed potential spaces, circulation paths, and architectural relationships. As the process developed, physical models were created using project, cardboard, threads, and structural elements to translate the sketches into three-dimensional forms. Each model became an experimental study of balance, proportion, enclosure, openness, and movement. Different materials were used to explore how forms interact with one another and how spatial experiences could evolve through layering, rotation, and intersection. The models were then photographed from multiple angles under different lighting conditions. Light and shadow studies became an important part of the exploration, as shadows revealed hidden forms, transitional spaces, depth, and circulation within the compositions. These photographs helped identify new spatial qualities that were not visible in the initial sketches or models. The extracted spaces and shadow compositions were translated back into sketches, where forms were refined further through architectural thinking. Rotated grids, intersecting planes, and fragmented geometries were organized into more coherent spatial systems. Through this process, abstract concepts gradually transformed into architectural forms and structured environments. In the final stage, the developed forms were converted into functional architectural planning. Plans, sections, and elevations were produced to define circulation, spatial hierarchy, public and private zones, and human interaction. The entire journey reflects a continuous transformation from abstract exploration to architectural realization, where sketches, models ,light, and spatial experimentation collectively shaped the final design outcome.

Authors

NameRegistration No.
Um e habiba2024-Arch-109
Project #50 Group 9

Velvet brutality

Supervisor Ar.Ahmed Iqbal and Aneeka Khan

Abstract

The project began with the psychological exploration of Hannibal lecter as the central character, chosen for his complex duality of sophistication and brutality. By deeply analyzing his personality, I extracted key descriptive words that represented his mental state, behavioral patterns, and hidden darkness. These keywords became the foundation for a series of abstract sketches, translating psychological traits into visual and spatial interpretations.From these abstractions, conceptual models were developed, allowing me to imagine inhabiting these forms as architectural spaces. Through sectional drawings, I began understanding how psychological dimensions could be transformed into spatial experiences. Using a nine-square grid, these sections were collaged into a structured framework, creating plans that simultaneously functioned as sectional representations of a dwelling.As the design evolved, the grid was refined to develop functional and symbolic spaces. The movement of the serial killer was then mapped within the three-dimensional cube using thread, introducing circulation as a narrative device. Further inspiration was drawn from Charles Rennie Mackintosh, whose spatial language informed a new interpretation of form and atmosphere. This was expanded through kaleidoscopic explorations, where fragmented perspectives revealed hidden spatial possibilities. A selected fragment from this kaleidoscopic study exposed wave-like geometries, leading to an investigation of wallproject groups and the tessellation principles of M.C. Escher. By integrating AI as a design tool, I merged Escher’s mathematical complexity with my extracted spatial fragment, generating multiple conceptual iterations. These AI-generated collages informed the creation of new planes and modular systems within the nine-square grid. Additionally, the spatial hierarchy and depth of Chand Baori (stepwell) were studied to enrich sectional complexity. Ultimately, the project was divided into two psychological zones—Hannibal tHe cannibal and His victim—within the grid framework. Through modular refinement and spatial synthesis, the final dwelling emerged as an architectural manifestation of psychological horror, narrative movement, and abstract transformation.

Authors

NameRegistration No.
Sehrish Mirza2024-ARCH-104
Project #51 Group 10

THE JOKER HOUSE

Supervisor Ar.Ahmed Iqbal and Aneeka Khan

Abstract

This project started with the idea of designing a house for the Joker, but during the process I realized that the project was not about creating a villain-themed house. It became an exploration of how architecture can represent human psychology, emotions, and fragmented identity through space, light, movement, and structure. In the beginning, I focused mainly on form and dramatic shapes, but gradually I understood that architecture is not only about appearance. Through research on Joker’s personality, trauma, loneliness, performance, and instability, I learned how spatial experiences can communicate emotions. Instead of designing normal rooms and furniture, I explored split levels, hidden circulation, ramps, reflections, monolithic materials, and fragmented spaces to create psychological transitions within the house. The project also taught me the importance of sections, sciography, structural logic, and spatial hierarchy. I learned how light and shadow can affect emotions, how movement can shape experience, and how architecture can become a narrative rather than just a building. By the end of the project, the house transformed into a psychological landscape where every space reflected a different emotional condition of the character. This process helped me understand architecture more deeply as an experience that connects human behavior, emotions, and space together.

Authors

NameRegistration No.
Seerat Touqeer2024-ARCH-126
Project #52 Group 11

House inspired by Eng.Mirza

Supervisor Ar.Ahmed Iqbal and Aneeka Khan

Abstract

Presenting a character-based studio project inspired by Engineer Muhammad Ali Mirza. My design reflects knowledge, discipline, storytelling, and spiritual awareness. Concept Line This space is not only designed for living, but for thinking, learning, and self-reflection. But to translate his personality into architecture. Character Explanation His personality combines logic and spirituality. That balance became the main idea of my design. About Space I used open and connected spaces to show clarity of thought and communication. The journey inside the project represents the journey of knowledge. Light and shadow are used symbolically to represent truth and understanding. Abstract Model These abstract forms are inspired by growth, movement, and intellectual depth. The model represents the transformation from confusion to clarity. Space Forming I created spaces that feel calm, focused, and meaningful. The forms are simple but emotionally strong. Planning The planning is functional but also symbolic. Each space has a purpose connected to the character’s lifestyle. Sections / Elevations The sections help express vertical movement and spiritual depth. The elevations are kept minimal to maintain simplicity and focus. Final Strong Ending This project is an architectural translation of personality, knowledge, and purpose. My aim was to create a space that inspires thinking, reflection, and character building.

Authors

NameRegistration No.
Soban Ahmed2024-Arch-118
Project #53 Group 12

Public Pavilion: An Exploration of Community, Space, and Interaction

Supervisor Mr.Haroon Malik

Abstract

This project explores the design of a public pavilion as a spatial structure that encourages social interaction, accessibility, and community engagement. Developed as an architectural study of form, openness, and human experience, the pavilion functions as a flexible public space that responds to its surrounding environment while creating opportunities for gathering, movement, and relaxation. The design investigates the relationship between structure and openness through the use of simple geometric forms and interconnected spatial elements. The pavilion is conceived as a lightweight and inviting enclosure that blurs the boundary between interior and exterior space. Through the arrangement of open pathways, shaded areas, and semi-enclosed zones, the project creates a dynamic spatial experience that accommodates both individual and collective activities. Natural light, ventilation, and circulation play a significant role in shaping the design. Openings and structural frames allow light and shadow to interact throughout the day, enhancing the atmosphere and visual character of the pavilion. The composition emphasizes balance, rhythm, and transparency, creating a sense of connectivity between users and the surrounding landscape. Through conceptual development and physical model-making, the project demonstrates how architectural interventions can activate public spaces and strengthen social interaction. The resulting pavilion serves as an exploration of contemporary public architecture where simplicity, functionality, and spatial experience combine to create an engaging and meaningful environment.

Authors

NameRegistration No.
Jassica Michelle2025-Arch-101
Anza Rauf2025-Arch-106
Ume Hafsa2025-Arch-105
Abdul Wasay Khan2025-Arch-118
Saymoon Wajid2025-Arch-103
Project #54 Group 13

Space Regeneration: Biomimicry Through Organic Architectural Form

Supervisor Mr.Haroon Malik

Abstract

This project explores the transformation of natural biological forms into architectural space through the concept of biomimicry. Inspired by the exoskeleton of the ground beetle, the design investigates how organic structures can inform spatial composition, structural efficiency, and architectural expression. The project develops a pavilion form derived from the beetle’s segmented shell, translating its curved geometry into a functional and experiential architectural enclosure. The design process begins with the observation and abstraction of the beetle’s physical characteristics, particularly its protective outer shell and rhythmic segmented structure. Through a sequence of geometric transformations, the biological form evolves into a series of interconnected curved surfaces that create shelter, enclosure, and spatial continuity. These shell-like segments generate an organic architectural language that balances structural stability with visual fluidity. The pavilion emphasizes the relationship between nature, movement, and space. Its curved form creates a dynamic interaction of light, shadow, and circulation while fostering a sense of protection and openness simultaneously. The project also explores how minimal structural elements can produce expressive spatial experiences through repetition, proportion, and form articulation. Through conceptual modeling and form exploration, Space. Regeneration demonstrates how natural systems can inspire innovative architectural solutions. The project highlights the potential of regenerative and nature-inspired design approaches in creating meaningful, adaptive, and sustainable spaces that strengthen the connection between the built environment and the natural world.

Authors

NameRegistration No.
Bisma Rashid2025-Arch-110
Abdul WasayKhan2025-Arch-118
Muzammil Ali2025-Arch-116
Aon Muhammad Abid2025-Arch-111
AMAMAH2025-Arch-104
Project #55 Group 14

Spatial Voids: An Exploration of Form, Light, and Negative

Supervisor Mr.Haroon Malik

Abstract

This project investigates the relationship between solid mass and void through the creation of an abstract architectural form. Developed as a study of spatial composition, the model begins with a simple cubic volume that is strategically carved to generate openings, recesses, and interconnected voids. These interventions transform a static geometric mass into a dynamic spatial experience, emphasizing the architectural significance of negative space. The design explores principles of balance, proportion, hierarchy, and visual permeability. Openings of varying sizes create a dialogue between interior and exterior spaces while allowing natural light to penetrate the form, producing changing patterns of light and shadow. The subtraction of volume not only reduces the perceived heaviness of the cube but also introduces depth, movement, and a sense of discovery within the composition. Through physical model-making, the project demonstrates how simple geometric manipulations can generate complex spatial qualities. The resulting form serves as an investigation into architectural expression, where voids become active design elements rather than mere absences of mass. Ultimately, the project highlights the potential of minimal interventions to create meaningful spatial experiences and contributes to a deeper understanding of form-making in architectural design.

Authors

NameRegistration No.
Haider Ali2025-Arch-113
Muzammil Ali2025-Arch-116
Ali Raza2025-Arch-107
Esha Fatima2025-Arch-114
Areeba Sajid2025-Arch-109
🧬

Biomedical Engineering

10 Projects

Project #56 Group 1

Development of an Accessible, Open-Source Extrusion-Based 3D Bioprinter for Hydrogel Scaffold Fabrication in Tissue Engineering

Supervisor Dr. M Rehan Ch

Abstract

Three dimensional bioprinting is an enabling technology in the form of an additive manufacturing process for constructing cell-laden structures that mimic the anatomy and function of native tissues. This project describes the design and fabrication of an inexpensive and open-source extrusion-based 3D bioprinter suitable for bioprinting of hydrogel scaffolds for applications such as tissue engineering, drug delivery and biomedical research. A three-axis (X, Y, Z) motion system is implemented with three NEMA-17 stepper motors and a framework which is manufactured using the fused deposition modelling method. A motion system which relies on an Arduino Mega 2560 R3 with a RAMPS 1.4 shield and TMC2209 stepper motor drivers is utilized to enable precise and quiet stepping movements. The design of scaffolds is developed using Fusion 360 CAD software and subsequently exported to G-code by using PrusaSlicer which is then transmitted to the 3D printer by Pronterface. 3 ml and 5 ml syringes can be utilized in the constructed extrusion mechanism and the system supports a build area of 200×200×300 mm. Feed rates and volumetric flow rates are adjusted in G-code in order to optimize print quality. Sodium alginate was chosen as the bioink for testing since it is a biodegradable, bio-compatible and commonly used material for bioprinting and its facile crosslinking can be easily incorporated into the printing system. This platform provides an affordable and easily reproducible system to researches for the development of scaffold based studies that can be applied to the tissue engineering, drug delivery systems or regenerative medicine

Authors

NameRegistration No.
Abdul muheet saqib2022-BM-129
Shayan khan2022-BME-120
Hafiz Fawad2022-BME-118
Project #57 Group 2

Mirror X Al powered smart mirror

Supervisor Dr. Umair Ahmad

Abstract

Modern lifestyles are becoming increasingly inactive, leading to a rise in lifestyle-related diseases. The high cost and limited access to professional personal trainers are barriers preventing people from working out on a regular basis. Current state-of-the-art smart mirrors only provide marginal improvements in health tracking, and are still limited by dependence on hardware sensors, lack of integration of physiological data, and poor implementation of real-time correction feedback. In this project report, we propose MirrorX which addresses such problems by integrating computer vision, real-time posture analysis and physiological measurement via multimodal methods in a single platform. MirrorX uses Google’s software suite MediaPipe to track 33 main skeletal landmarks in the human body to determine the angle of joints when doing ten specified exercises including squat, lunge, push-ups, plank and biceps curl. The calculated angles of the joints are compared with the biomechanical criteria to check the correctness of the posture. Users have a convenient web interface integrated in the mirror itself to get real-time feedback for improvement. The MirrorX is also equipped with AD8232 sensors for ECG-based heart rate calculation, MyoWare 2.0 for EMG sensing and estimation of muscle exhaustion level and MH-Z19C for measuring CO2 concentration in exhaled air. The ESP32 microcontroller transmits all the physiological signals to the wireless. Such integration of pose estimation and intelligent repetition count as well as physiological parameter measurement provides an opportunity for building a smart fitness coaching system suitable both for home and clinical use in physical therapy sessions , contributing to the development of intelligent health care solutions and ambient assisted living

Authors

NameRegistration No.
Khadija Rafique2022-BME-111
Ume Areeba2022-BME-124
Project #58 Group 3

oxygen therapy device for ulcers

Supervisor Dr. Umair Ahmad

Abstract

Diabetic foot ulcers (DFUs) remain one of the most prevalent and clinically burdensome complications of diabetes mellitus, affecting up to 25% of diabetic patients over their lifetime and representing a leading cause of non-traumatic lower-limb amputation worldwide. The principal pathophysiological barrier to DFU healing is chronic tissue hypoxia, arising from peripheral angiopathy and neuropathy, which impairs the cellular mechanisms of immune defence, collagen synthesis, and angiogenesis essential for wound repair. Existing topical oxygen therapy (TOT) devices, whilst clinically validated, are prohibitively expensive and largely inaccessible in resource-limited healthcare settings, creating a critical gap between established therapeutic evidence and real-world clinical deployment. In this project, we propose an Arduino-based portable oxygen therapy device which addresses these limitations by integrating closed-loop flow control, real-time physiological monitoring, and a user-adjustable delivery interface within a single low-cost hardware platform. It employs an Arduino UNO microcontroller as its central processing unit to execute a pulse-width modulation (PWM) control algorithm that regulates a 12 V DC diaphragm pump across a therapeutic flow range of L/min. A Hall-effect YF-S201 flow sensor continuously monitors oxygen delivery rate, while a DS18B20 digital temperature sensor tracks the thermal condition of the gas pathway — a parameter directly relevant to warm clinical literature. A PC817 optocoupler and IRF244N power MOSFET oxygen delivery protocols documented in the form the high-side switching circuit, with power conditioning provided by an LM2596 step-down voltage regulator. Users interact with the device through a TFT LCD on which there are three options, or operator decide the option of gas flow according to the patient health.The full circuit was designed and virtually validated in Proteus Design Suite 8.15 prior to breadboard prototyping and PCB fabrication. Such integration of microcontroller-based closed-loop regulation, multi-parameter physiological sensing, and affordable hardware demonstrates the feasibility of developing clinical-grade oxygen therapy technology suitable for deployment in low-resource wound care settings, contributing to the broader goals of accessible biomedical engineering and equitable healthcare solutions

Authors

NameRegistration No.
Mahnoor Majeed2022-BME-123
Muhamad Hamza2022-BME103
Project #59 Group 4

A Portable UVC-LED Sterilization System for Stethoscope Decontamination in Clinical Settings

Supervisor Dr. Sameen Ahmed

Abstract

In this context, the use of such tools as stethoscopes may be associated with the spreading of dangerous bacteria between different patients because of their improper disinfection, which causes the occurrence of healthcare-associated infections (HAIs). Therefore, the aim of this research is the prevention of stethoscope diaphragm contamination with bacteria by the means of disinfection with the help of UV-C radiation without any chemicals. The purpose of the study was to find out the influence of various wavelengths of UV-C radiation and their duration of action on the development of germs that may appear on the surface of the stethoscope that was infected by them. Thus, four wavelengths of UV-C LED, such as 260 nm, 265 nm, 270 nm, and 275 nm, were used. Escherichia coli and Staphylococcus aureus, two types of bacteria which often cause health care-associated infections, were selected for experimentations. Samples infected with these bacteria were then exposed to UV-C rays for different periods such as 5, 10, 15, 20, 25, and 30 seconds on agar plates. The bacterial samples were streaked on the nutrient agar plates and left for incubation after exposure of different time periods. Efficacy of the disinfection process was evaluated through the comparison of the number of colonies formed after exposure to UV-C compared to the control sample. The results obtained from the experiment showed that an increase in exposure time caused UV-C rays to effectively reduce bacterial formation. Effectiveness of UV-C LED as a means of reducing contamination of stethoscope surface through bacteria formation was shown as increased exposure time led to a reduced number of bacterial colonies. Comparative analysis of effectiveness of different UV-C rays for inactivating bacteria also became simpler because of the comparative analysis performed. It has been clearly shown in the present study that there is great potential in using the UV-C LED technology as a cheap and efficient disinfection method for non-critical medical devices.

Authors

NameRegistration No.
Arooj Fatima2022-BME-117
Nimra Ejaz2022-BME128
Saba Rani2022-BME-116
Project #60 Group 5

Machine Learning-Enhanced Near-Infrared Spectroscopy System for Non-Invasive Tissue Oxygenation Monitoring

Supervisor Dr. Salman Ajmal

Abstract

This project presents a machine learning-enhanced near-infrared spectroscopy system for non-invasive tissue oxygenation monitoring. Tissue hypoxia is an important problem in clinical settings, since conventional pulse oximeters are limited to measuring arterial oxygen saturation from peripheral sites and cannot detect tissue-level deterioration that may precede any change in SpO₂. Current near-infrared spectroscopy (NIRS)-based systems are expensive, bulky, and hospital-bound, making them unavailable in low-resource environments. In this work, we tackle this problem by proposing a low-cost, portable system that integrates pulse oximetry with tissue-level optical sensing and machine learning analysis. The system consists of a microcontroller, optical sensors for measuring SpO₂ and heart rate, a motion sensor for artifact detection, a display for real-time output, and an SD card for data logging. For tissue monitoring, a near-infrared LED and photodetector capture deeper tissue optical variations. The collected physiological signals are assembled into a structured dataset for preprocessing, feature extraction, and future machine learning-based analysis aimed at identifying oxygenation patterns, reducing motion-related artifacts, and supporting intelligent physiological assessment. This prototype platform demonstrates a potential pathway towards affordable tissue oxygenation monitoring for applications such as low-resource clinical settings, sports physiology, and remote patient care.

Authors

NameRegistration No.
Maria Fatima2022-BME-112
Zartash Anjum2022-BME-109
Saira Sajid2022-BME119
Project #61 Group 6

Smart Multi-Parameter Health Monitoring System Iot with blynk app

Supervisor Dr. Umair Ahmad

Abstract

In this project, a Smart Multi-Parameter Health Monitoring System using ESP32 as the main hardware component along with other IoT technologies is developed for the monitoring of various physiological and environmental factors. This health monitoring system has been developed in such a manner that it provides continuous monitoring of multiple physiological and environmental parameters using only one portable monitoring device. Various biometric sensors have been used in the proposed health monitoring system which include MAX30102 for heart rate and SpO2 measurement, ECG sensor to detect electrical signals of the heart, DS18B20 as a body temperature sensor, SCD30 to sense the levels of CO2 in environment and end tidal, humidity, and ambient temperature, and pH analog sensor for pH measurement. ESP32 captures information from all connected sensors, then processes the data obtained from these sensors, and sends them via Wi-Fi to the Blynk IoT platform from where the users can monitor the parameters through the Blynk app on their smartphones. The developed prototype of Smart Health Monitoring System is inexpensive, portable, and user-friendly; thus, it is very useful for home healthcare monitoring, remote monitoring of patients, biomedical researches, and preventive healthcare measures.

Authors

NameRegistration No.
Amna Amjad2022-BME-130
Kaneez Fatima2022-BME-110
Project #62 Group 7

A Hybrid MODWT-Based PQRST Feature Extraction and Dual-Branch Deep Learning Framework for Interpretable Cardiovascular Disease Classification on MIMIC-IV-ECG

Supervisor Dr. M Rehan Ch

Abstract

There is a trade-off between diagnostic accuracy provided by deep learning models and interpretability in the context of automated ECG classification in the cardiovascular diseases domain. This research proposes an integrated two-phase framework that addresses the problem via the use of a clinically validated classical digital signal processing front-end coupled with a deep neural network classifier, both validated on the MIMIC-IV-ECG database with 800,035 records. In the first phase, a QRS complex detection algorithm based on the Pan-Tompkins algorithm and maximal overlap discrete wavelet transform multi-resolution analysis with a symlet-4 basis is used to perform heartbeat delineation and extract a 13-dimensional clinically interpretable set of features per heartbeat, including heart rate, QRS width, QTc interval, ST segment deviation, P wave duration and amplitude, T wave amplitude, R wave amplitude, RR standard deviation, PR interval, Q wave amplitude, QRS front plane axis, and P wave morphology score. Classification thresholds for five super-classes of cardiovascular diseases – rhythm disturbances, conduction problems, hypertrophy, repolarization disturbances, and ectopic heartbeats – are generated using P5/P95 percentiles of 701,163 MIMIC normal records.During the second step, a two-branch neural classifier combines the 13-dimensional feature vector, which can be interpreted by clinicians, with the learned representation of morphological patterns derived from a 1D convolutional neural network processing windowed raw ECGs segmented at the level of beats, thereby allowing the use of both domain-informed hand-crafted features and machine-learned latent features. The proposed hybrid framework is trained and evaluated using the MIMIC-IV-ECG dataset through the optimization of multi-label binary cross-entropy loss since the medical truth entails that many different cardiac diseases coexist in the same patient. The feature extraction process is assessed against machine-extracted ground truth, yielding 0.0% heart rate error, 3.5% QT interval error, and 11.6% QRS duration error. Classification performance is measured according to sensitivity, specificity, and F1-score on a per super-class basis, comparing our approach to both the rules-based baseline and the pure 1D-CNN. Our proposed framework highlights that MODWT-derived PQRST features contain discriminatory features different from those contained in raw waveforms' CNN features, which results in increased multi-label classification performance while maintaining interpretability.

Authors

NameRegistration No.
Ali Murrtaza2023-CD-BME-140
Mariam2022-BME-102
Umaima Shakeel2022-BME-114
Project #63 Group 8

Smart Rehabilitation System for upper Limbs Recovery Using Electrical Muscle Stimulation

Supervisor Dr. Salman Ajmal

Abstract

An Upper limb Motor Impairment is a clinical condition in which a person experiences loss or reduction of voluntary elbow movement which arises as a consequence of stroke or peripheral nerve injury. When Somone has this problem everyday tasks become very difficult things like eating, getting dressed or picking up objects. This makes it hard for them to live independently and affects their overall quality of life. Electrical Muscle Stimulation is a treatment that uses gentle electric signals to help the muscle work again and it has been shown to work well in this condition However, the EMS device currently available in market has some big drawbacks. They are too long, too complicated and not easy to use at home. They also can’t be adjusted to meet the specific need for each patient. This project presents the design of Smart Rehabilitation System developed to support recovery of elbow joint movement particularly flexion and extension. The system targets the two primary muscle responsible for Biceps brachii and Triceps Brachii by delivering controlled Electrical Stimulation. A microcontroller circuit is used to generate safe and accurate stimulation pulses which allows therapy parameters such as intensity, frequency and duration to be adjusted according to the individual needs of the patient. The proposed rehabilitation system provides a simple and user-friendly solution for stroke recovery. The system is designed to make rehabilitation easier by automatically organizing therapy sessions and it adjusts stimulation levels accordingly. This provides smooth transition that improves the patient comfort and prevents sudden muscle contractions. It also helps to reduce muscle fatigue gradually. It was observed that the patients were able to bend and straighten their elbows more smoothly and with better control. They were able to move their arms with wider range than before. The Easy to use -interface made the system easier to set and operate and reduces the complexity often related with traditional EMS devices. This system as a whole provides a practical, compact and efficient rehabilitation solution that can help stroke patients continue their recovery comfortably at home. Keywords: Stroke Rehabilitation , Electrical Stimulation System(EMS) , Upper limb Impairment, Control Stimulation therapy

Authors

NameRegistration No.
Baha2022-BME-106
Arslan Jillani2022-BME-107
M Nadeem Anwar2022-BME-115
Project #64 Group 9

To find the association of Inflammation-Linked microRNAs (miR-146a & miR-155) with its target genes in Dry Eye Disease (DED): in silico Analysis

Supervisor Dr. Bisma Rauff

Abstract

Dry Eye Disease (DED) is a multifactorial ocular surface disorder characterized by tear film instability, hyperosmolarity, inflammation, and ocular surface damage. Dry eye disease results in eye irritation, burning sensation, vision impairment, and eye pain, significantly impacting the quality of life of many millions of people around the world. There are several factors responsible for the increase in prevalence of dry eye disease, such as the surrounding environment, the use of electronic devices, aging, hormonal dysfunction, and autoimmune disorders. Several recent studies have shown that inflammation is an important risk factor in dry eye disease development through immune signal pathway activation and secretion of pro-inflammatory cytokines like IL-1β, IL-6, and TNF-α. The microRNAs (miRNAs) have gained attention lately for their role in controlling immune and inflammatory responses post-transcriptionally. From among all the miRNAs, miR-146a and miR-155 have emerged as critical miRNAs associated with inflammation. Both miRNAs play an essential role in immune cell activation and cytokine signaling, among other inflammatory processes, in various inflammatory and autoimmune diseases. However, there is little literature on their joint target gene interaction network in Dry Eye Disease. In the present study, in-silico bioinformatics approaches have been employed to explore the connection between miR-146a and miR-155 and their targets in Dry Eye Disease. Firstly, the target genes of the two miRNAs will be obtained using authenticated miRNA databases; then, common targets will be identified. For the PPI network analysis, STRING database will be used, and hub genes will be selected using Cytoscape and CytoHubba plugins. More biological pathways associated with the pathogenesis of dry eye disease can be elucidated through functional enrichment analysis. It is likely that possible inflammatory biomarkers and treatment targets for future microRNA-based DED diagnosis and treatment strategies could emerge from this study

Authors

NameRegistration No.
Aqsa Akram2022-BME-122
Malik M Mudassir2022-BME-104
Muhamad Jahanzaib2022-BME-113
Project #65 Group 10

A Portable UVC-LED Sterilization System for Ultrasound Probe Decontamination in Clinical Settings

Supervisor Dr. Sameen Ahmed

Abstract

Healthcare-associated infections (HAIs) continue to be a major problem in today's healthcare facilities, especially because of inadequate disinfection of reusable medical devices. Ultrasound probes are common diagnostic devices that frequently contact patients’ skin and mucous membranes and require rigorous disinfection to prevent cross-contamination. Traditional chemical sterilization techniques can be slow and operator dependent, and may leave chemical residues or damage sensitive equipment over time. Recent developments in ultraviolet-C (UVC) lightemitting diode (LED) technology have provided a promising alternative for fast and chemical-free sterilization. This project presents the design and conceptual development of a portable sterilization system based on UVC for ultrasound probe decontamination. The suggested system can effectively inactivate microorganisms by means of high-power UVC LEDs operating at the germicidal wavelength range of approximately 265–280 nm According to research in the literature, UVC radiation successfully inactivates bacteria by causing damage to their DNA and RNA structures, which stops them from replicating. Also focuses system design, methodology, microbiological validation principles and engineering considerations for effective sterilization. The study also reviews the current experimental and clinical studies on the effectiveness of UVC sterilization, microbiological testing techniques, and the performance benefits of UVC LEDs. Common bacterial indicators such as Escherichia coli and Staphylococcus aureus are relevant to the evaluation of the efficacy of sterilization by agar-based colony count methods. This project shows that portable UVC-LED sterilization systems can be affordable, environmentally friendly, and suitable for clinical use. They may improve infection control and ultrasound probe decontamination practices in healthcare settings.

Authors

NameRegistration No.
Hina Rauff2022-BME-121
Amisha Khan2022-BME-126