مقترحات مشاريع التخرج

The integration of Machine Learning (ML) techniques in Internet/Network Intrusion Detection (IDS) systems create efficient and high-performance entity in detecting anomalies within IIoT systems. To train the IDS, well-known datasets found on the IEEE data port website are professional sets and used by students, researchers, and others to test existing ML as well as new ones. The work in this project require the student to use Google Colab and existing cybersecurity datasets to train IDS using Python.

The advancement in separating the control plane from the data plane enables decision making through centralized controllers that guide network switches (packet forwarding) to implement different network operations. This is known as SDN (software defined networking) and is limited to the predefined capabilities of the switch’s firmware or ASIC. However, if we need to implement packet operation, modify headers, customizing protocol then we need to redefine the switch’s behavior at the packet level. This is what the project about and the student will use the P4 programming language in Mininet environment and BMv2 switch to achieve predefined packet processing tasks.

------------------------

The usage of wireless sensor networks (WSN) is widely conducted in many applications such as monitoring and smart cities. However, an issue of limited energy still requires a routing protocol that help to prolong the sensor lifetime. This project aims to compare and evaluate the performance of two well-known protocols LEACH (Low-Energy Adaptive clustering Hierarchy ) and TEEN (Threshold sensitive Energy Efficient sensor Network ). Using MATLAB simulation tool, both protocols will be analyzed and evaluated under similar network conditions to measure the Energy consumptions, packet delivery ratio, and delay. These results will provide insights into the strength and weakness es of both protocols to help in selecting a suitable scenario in the application.

This study develops a cloud-based management system for training courses at Al-Nahrain University. The system enhances course organization, registration, and monitoring by providing manageable, scalable, and accessible services. It reduces administrative effort, improves user experience, and supports efficient decision-making.

This project aims to simulate an AI-powered Internet of Things (IoT) network. The network will be designed to monitor number of sensors data, detect anomalies, and make autonomous decisions. The system simulates (using python 3.8) a set of sensors generate real-time data streams—such as temperature, humidity, or motion. The readings are then transmitted (using MQTT protocol) to a centralized server in predefined time intervals. The server has a simple AI anomaly detection module where the incoming data analyzed (using Isolation Forest) to identify abnormal or unexpected readings. Upon detecting an anomaly, the AI decides whether to discard the faulty record (e.g., due to sensor error) or trigger a predefined response (issuing alerts). N.B. The simulation can be implemented via existed (ready to use functions) can be found on the web. The student need to understand the general structure of the network and how the different parts are connected to each other’s. Moreover, the student need to understand the job of each element not the details of their operation.


The focus for the student is to:

• Understand the overall network structure.
• Learn how the different components (sensors, MQTT, server, AI module) connect and interact.
• Understand the role of each component, without going too deep into technical details.

Software-Defined Networking (SDN) is an approach to manage networks through the decoupling the control plane from the data plane, and using a centralized / programmable controller. Even though SDN offers an improved, flexible and efficient operation, this come with new security challenges due to the centralized architecture. This project presents a framework for detecting and mitigating network attacks in an SDN environment using a custom security application deployed on an SDN controller. The proposed solution uses the Ryu controller to monitor, analyze, and detect abnormal patterns in network traffic that indicate common network attacks such as TCP SYN Flood Attack. Upon detection, the controller dynamically applies one of the mitigation strategies to contain the attack and maintain network performance. The test of the system is for both of detection accuracy and response time. The system is created via mininet. To monitor traffic Python + Ryu app will be used. The detection and mitigation will use SYN flood detection logic and drop rules respectively.


Steps

• Use Mininet to simulate the SDN network.
• Install and configure the Ryu controller.
• Develop a Python application on the Ryu controller.
• Continuously monitor and analyze incoming network traffic.
• Implement SYN Flood detection logic to identify abnormal traffic patterns.
• When an attack is detected:
  
  • Apply dynamic drop rules to block malicious traffic.
  • Maintain normal network performance.
• Measure detection accuracy (how well attacks are caught).
• Measure response time (how fast the system reacts).

Private cloud services comes with a cost, especially when cloud providers increase the subscription fees over the time. Small businesses, such as, shops, clinics, etc. need to overcome such high costs when dealing with proportionally simple cloud services. This project is intended to help such businesses with private cloud services that fulfil both lower cost and privacy, using well-known Raspberry PI 5 device. The personal cloud server can feature customizable and configurable services tuned for the business, which might not available when using private cloud providers. Finally, the project will be tested for services and feasibility and compared with common cloud service providers.

The use of electrical and plug-in hybrid automobile is increasing dramatically every day. Charging stations will spread everywhere and might replace gas stations eventually. Monitoring charging stations is a crucial task, since it is expected that charging stations will rely on sustainable energy, such as solar power, instead of using national power grid. This category of stations depends mainly on giant batteries that store the energy when sustainable energy is available. Therefore, there is a limited energy capacity for each station, similarly to gas stations which is limited to the capacity of its tanks. The main goal of this project is to monitor charging stations using IoT devices and publish the available energy levels for each station online, such that when an electrical vehicle needs to recharge it can learn if the intended station has enough energy, in terms of Kilo Watts (Kw). The project will also introduce a web application or a mobile app that can suggest nearby charging stations to vehicles with its specifications such as: the amount of available energy, and its price per Kw.

This project develops a smart IoT-based medicine reminder system enhanced with an AI chatbot interface. The system tracks medication schedules using IoT sensors (reed switches or weight sensors in a pillbox) and alerts patients when it’s time to take their medicine. The chatbot interacts with the user via Telegram, or WhatsApp, providing reminders, logging adherence, answering questions about medication schedules, and sending notifications to caregivers when doses are missed.

This project aims to design and implement a smart environmental monitoring system using IoT technology. The system will measure temperature, humidity, and gas levels using sensors connected to an ESP32 microcontroller. The collected data will be transmitted in real-time to a cloud database (Firebase) and displayed on a web dashboard with alert notifications when threshold values are exceeded. The proposed solution is low-cost, scalable, and can be extended to applications such as smart cities and smart agriculture.

With the growing popularity of football globally, the accuracy of each and every events and technology involved in the game is a concern. Critical situations arise when the referee cannot discriminate a goal or no goal by fine margins due to human visual limitations. Nowadays Video-Assistant Referee (VAR) and other technologies performing accurate decision-making can be implemented in a live match. But the process consumes a lot of time which reduces the fast pace of the game and can cause unnecessary distractions. This study aims to design an automatic goal-line detection system with the help of radio frequency identification (RFID) – Arduino interfacing. RFID incorporates the use of radio waves to extract the information stored in a tag attached to an object. The proposed system uses RFID tags which are fitted on the inside surface of a football. The information embedded in the tags is read by RFID readers, placed behind the goalpost and directly feedback to the referee. This technology does not require any additional programming and camera analysis in decision-making, thus making the system faster to help the referees in quick decision-making and maintaining the pace of the game.

This project integrates Internet of Things (IoT) sensors with a cloud-based platform to provide real-time visibility into the movement and condition of goods within a supply chain. It aims to enhance efficiency, transparency, and product quality from the point of origin to the final destination. The system works as follows:
• IoT Devices: Sensors (GPS, temperature, humidity, and shock) are attached to shipments to collect data. These sensors are connected to microcontrollers (like Arduino or Raspberry Pi) that process the information.
• Cloud Platform: A central hub that receives, stores, and processes the data from the IoT devices. This platform includes a database, servers, APIs, and a user-friendly dashboard for data visualization.
• User Application: A web or mobile application that allows users (logistics managers, companies, or customers) to access real-time information and track their shipments.
This system provides full transparency and predictive insights, reducing errors and delays. It ensures product integrity, especially for sensitive goods, and significantly improves overall customer satisfaction by allowing them to track their orders effortlessly.

Wireless Sensor Networks (WSNs) are widely used in applications requiring large-scale, energy-constrained monitoring, such as environmental sensing and industrial automation. However, efficient energy usage remains a critical challenge. This project introduces an AI-enhanced clustering protocol that integrates Particle Swarm Optimization (PSO) and Artificial Neural Networks (ANN) into the conventional LEACH (Low-Energy Adaptive Clustering Hierarchy) framework. PSO is employed to generate optimal cluster head candidates, while the ANN dynamically evaluates these candidates based on features such as residual energy, distance to the sink, and round number. The ANN guides the final selection of cluster heads to achieve more balanced energy distribution and extended network lifetime. Simulation results in MATLAB demonstrate that the proposed approach significantly outperforms traditional LEACH and PSO-based methods in terms of node survival and energy efficiency, providing a scalable and intelligent solution for modern WSN

This project presents the design and implementation of a Virtual Mini Data Center intended as an educational framework for university laboratories and computer networking courses. The environment provides essential network services — including a File Server, Web Server, DNS Server, DHCP Server, and Directory Services — all deployed in an isolated virtual network that simulates real-world data center functionality. The primary goal is to deliver a step-by-step lab guide that enables students to deploy, configure, and manage these services in a reproducible manner. The guide includes clear instructions, commands, and screenshots to support both instructors and learners. For practical implementation, the project demonstrates how these services can be set up on Linux-based virtual machines using Samba/NFS for file sharing, Apache/Nginx for web hosting, BIND for DNS, ISC DHCP for IP leasing, and Samba 4 for Directory Services.

With the rapid increase of cyber-attacks targeting modern networks, there is a growing need for effective mechanisms to detect and prevent malicious activities. This project aims to design and implement an Intrusion Detection System (IDS) using the open-source tool Snort in order to evaluate its effectiveness in detecting common network threats such as Denial-of-Service (DoS) attacks, port scanning, and intrusion attempts. A virtualized network environment will be built to simulate real-world scenarios, where attacks are generated using penetration testing tools such as Nmap and Metasploit, while Snort monitors the traffic and generates alerts upon detecting suspicious activities. The performance of the IDS will be analyzed through key metrics including detection rate, false alarm rate, and response time.

The project is a web application that provide essential administrative tasks for academic department rapporteurs. This system manages meeting minutes, basic report generation, and departmental record-keeping through a UI/UX portal, while demonstrating fundamental network authentication concepts. By implementing a simple virtualized network environment, the system will provide basic security practices including IP-based access control and user authentication tied to network location. The project prioritizes core functionality over advanced features, focusing on document management capabilities that would enhance and assist department rapporteurs to increase productivity and reduce time.

Security, confidentiality, and availability of examination records are important issues to academic institutions. Existing manual workflow in examination committees leads to multiple avenues of information security threats, unauthorized access, and data manipulation. This project involves the design and implementation of a Secure Exam Committee Management System hosted in a secure network.

This project presents the design and development of an IoT-enabled security system utilizing Passive Infrared (PIR) technology for motion detection. The system employs a PIR sensor to detect human movement by sensing variations in infrared radiation emitted from the body. Upon detecting motion, the system communicates through a Wi-Fi-enabled microcontroller to send instant notifications to the user via email. This integration of PIR sensing with IoT ensures real-time monitoring, low power consumption, and cost-effectiveness, making it suitable for smart homes, offices, and industrial security applications.

Current lower-limb prosthetic systems face limitations in delivering timely, adaptive feedback during real-world use. This often reduces the effectiveness of rehabilitation and leaves both users and clinicians without actionable, real-time insights. This project proposes the design and implementation of an intelligent lower-limb prosthetic system that integrates IoT-based sensor networks, real-time data communication, AI-driven gait analysis, and immersive VR feedback. The prosthetic device will be equipped with embedded sensors (e.g., pressure, motion, and orientation) connected to a wireless microcontroller (ESP32), which transmits biomechanical data over a secure, low-latency IoT infrastructure.The transmitted data will be analyzed using machine learning models to identify gait patterns, detect irregularities, and provide corrective feedback. This feedback will be delivered visually via a virtual reality interface, allowing the user to observe and adjust movement in an interactive and adaptive environment. A mobile application will be developed to serve as an access point for monitoring data, visualizing analytics, and configuring system parameters.The project focuses on designing a low-latency, fault-tolerant communication layer between wearable edge nodes and distributed analytical components. From a clinical perspective, it is expected to reduce feedback response time and enhance rehabilitation engagement through immersive, user-specific guidance. Unlike conventional prosthetic systems, this approach aims to bridge real-time gait analysis with VR- based corrective feedback by uniquely integrating IoT, AI, and VR into a unified, interactive rehabilitation platform. If successfully implemented, the system could accelerate patient recovery and improve long-term mobility outcomes, contributing to enhanced quality of life and reduced clinical burden.

This project deploys a multi-protocol honeypot (e.g., Cowrie/OpenCanary) to attract and study common attacks such as brute-force SSH, web scans, and SMB probes. Collected events are shipped to a central index for analysis and daily roll-ups. The student produces a lightweight threat-intelligence report for the local environment, including source ASNs, credential patterns, and observed TTPs, plus practical hardening recommendations for small organizations .

This project implements a hardened enterprise resolver with DNS-over-TLS/DNS-over-HTTPS, optional DNSSEC validation, and threat filtering via curated blocklists. Using Unbound or Bind plus Pi-hole, the student delivers privacy and security for client queries while generating useful observability (dashboards of query types, NXDOMAIN spikes, and blocked domains). Evaluation compares baseline vs hardened setups for latency, cache hit ratio, and blocked-threat coverage. Deliverables: configs, blocklist strategy, dashboards, and a before/after report.