1. Intelligent Driver Behavior Detection Systems Supervised by Prof Jules R Tapamo
In this project I designed and implemented a driver monitoring system to monitor two major causes of road accidents, drowsiness and intoxication. The drowsy detector used a video web camera that pointed directly towards the driver’s face and monitored the driver’s eyes to determine fatigue.
The first step in the drowsy detection was the acquisition of the driver’s image and each frame from the video stream. Face and eye detection followed using methods based on the Viola-Jones algorithm. The eye pair of the driver was extracted as the region of interest (ROI) and passed into a trained neural network model for prediction on whether the eyes are closed or opened. If the eyes are found to be closed for a period longer than the threshold the system produces an audio alert informing the driver to wake up.
The neural network model was based on 40*40 (size = 1600) images of closed and opened eye pairs, has two outputs either class 0 for closed eyes or class 1 for opened eyes and 20 hidden layers. This part of the system worked under reasonable lighting conditions.
The alcohol detector is based on an alcohol sensor MQ-3 that detects the driver’s alcohol level, if the alcohol level is above the limit the systems warns the driver not to drive in this state and sends a message to his/her relatives using the GSM technology. The warning is audio based and this part of the system worked under all conditions.
Skills acquired: Software and hardware debugging
OpenCV for image processing
Embedded C for ATMEGA 8 Microcontrollers
Python programming for interfacing hardware and software
Time management and planning
AT commands and SIM900 GSM/GPS modem
2. Wirelessly Controlled Wall Climbing Robot Supervised by Dr Tom Walingo
In this project a wireless controlled wall climbing robot was designed by a group of four members. The motivation behind the design was to make it possible to access places that could be otherwise difficult or dangerous for human operators to access and also to eliminate the erection of costly scaffolds in the construction industry.
Skills acquired: Feasibility studies and background research
Team work and team leadership
Microsoft Visual Basic
Report writing and oral presentation skills
Software and hardware debugging
Printed Circuit Boards (PCB) design
3. Stop and Wait Automatic Repeat Request Simulation (SAW ARQ) supervised by Mr. P.N Zulu
Automatic repeat request is one of the mechanisms for error handling in data communications and is used to provide reliable data transfer. Stop and wait is the simplest of ARQ schemes as the receiver and transmitter work on one frame at a time. In this project the stop and wait protocol was investigated and the simulation was implemented using C++.
Skills acquired: Data Communications
4. Implementation of the Data Encryption Standard supervised by Dr Colin Chibaya
In this project algorithms were designed, analyzed and eventually used in the implementation of the Data Encryption Standard (DES). The symmetric-key algorithm (DES) was predominately used for the encryption of electronic data was implemented using Microsoft Visual Basic.
Understanding of: Encryption of data
Advanced Encryption Standard (AES)