dissertation.md

# List of Dissertation Topics

The new autonomous Vehicles development group are looking for a number of highly skilled and motivated final year students to work on a number of high profile research topics. Please read the list of topics below and, if you are interested email m.tyers@coventry.ac.uk attaching your CV and explaining how you would go about carrying out the work.

## Overview

Below are a list of potential dissertation topics. These are not for the faint-hearted! Each will require a high level of dedication but will enable you to become an expert in a given field and, if successful you will be invited to be part of the new Autonomous Vehicles development team called **AI Coventry**.

### Learn – Develop – Win

Coventry University is fully committed to the training and development of the next generation of autonomous vehicle engineers. This is a fiercely competitive environment and, to help stimulate our students to give their best, we enter them into some of the toughest competitions available to Universities.
The best way to learn is to be placed in a competitive environment and, in all the dissertation topics, this is what will happen. This is the Coventry University Way...

### Formula Student – AI

The IMechE have launched an Autonomous Racing Car competition that runs alongside the main Formula Student event.  This runs at the Silverstone Grand Prix Circuit in the week after the British Grand Prix. This year (2019) we won all the autonomous races to place 2nd overall. All the autonomous car research on the following pages is designed to prepare for next year's competition and, if you take on one of these research challenges (and succeed) you will be invited to be part of next year's team and compete at Silverstone.

### WRSC

Coventry University also competes in several autonomous sailing boat competitions, including the European RoboBoat challenge (held in the Baltic in May) and the global WRSC (held in China this year). Next year we also plan on entering the SailBot competition in the USA!

### How to Get Involved

We are looking for students to get involved in all these competitions. If you are interested you should choose one of these dissertation topics. If you do this and succeed you will be eligible to be part of the team. If you are interested, email me (m.tyers@coventry.ac.uk) and let me know which project(s) you are interested in working on.

## Differential GPS

Standard GPS has inherent inaccuracies, typically +-80cm. [DGPS](https://racelogic.support/01VBOX_Automotive/01General_Information/Knowledge_Base/How_Does_DGPS_(Differential_GPS)_Work%3F) is a new type of system to provide positional corrections to GPS signals and thereby achieve accuracies of +-2cm. Developing such a system is absolutely critical to the success of all forms of autonomous vehicles.

Current off-the-shelf systems typically run to £around £2K which means their applications are limited. The project would be to develop and evaluate a low-cost DGPS system for use on the full range of autonomous vehicles being developed.

## Odometry, Logging and Telemetry

A key tool used by motor racing teams is the capture and live analysis of sensor data taken from the moving vehicle. The current autonomous car has rudimentary logging to the hard drive however for next year we want to expand this so we can fully understand the vehicle dynamics.

We are looking for a student to install sensors to capture important dynamic data and to develop a telemetry system to enable this data to be sent live to an analysis tool. If successful the student would be invited to join the FS-AI 2020 team and compete at Silverstone.

All our autonomous vehicles require some form of live telemetry data to allow the developers to understand performance in real time. Given the distances involved, BlueTooth and WiFi are not sufficient. One recent technology is a type of long-range WiFI called LoRaWAN. This project is to develop a telemetry network capable of transmitting real-time data at ranges of up to 10kM.

## Stereoscopic Computer Vision

This year's competition car used a Z-Camera to provide 3D vision which was used to navigate around a coned course. Due to the large amount of data being generated hitting the limits of USB3 it was limited to 1080p at 30fps. Depth data was not precise since the baseline between the cameras was small. If the next car is to travel at a higher speed we need to both increase depth data accuracy whilst increasing the framerate.
This project would require the development of a custom stereoscopic camera system with adjustable baseline. Each camera would have its own dedicated computer, possibly a Jetson Nano for edge processing of data to reduce the volume of data per frame and the two processed feeds would be passed to a third computer for depth data generation.

If successful the student would be invited to join the FS-AI 2020 team and compete at Silverstone.

## Track Following and Obstacle Avoidance

The FS-AI competition 2019 focussed on competing on tracks delineated by coloured cones. Whilst we don't know much about next year's competition we do know they will be a lot more challenging and will require the car to race around standard tracks. For this to work we need to develop algorithms that can detect and map the track and avoid any obstacles placed in the path of the vehicle.
We are looking for a student to develop software, possibly based on ROS to detect the edges of the track and use this to both map the track and guide the car around. If successful the student would be invited to join the FS-AI 2020 team and compete at Silverstone.

## School Autonomous Car Challenge

The IMechE and IET have made it a priority to encourage more students to become autonomous vehicle engineers and are keen to get school children interested in this. Last year a group of stage 2 CS students designed and built a low-cost autonomous car using an RC Car chassis and a raspberry Pi which encouraged them to take the subject further and get involved in the FS-AI competition.

We are looking for a student to re-engineer the existing prototype into a teaching tool that will introduce school children to the concepts around designing autonomous cars. In order to make this as accessible as possible the system could be based on the BBC Micro-Bit. Whilst the development of the vehicle is important the project should also involve the development of a set of teaching resources, these can be tested on the visiting students during the open days as well as in local secondary schools. A mentor from one of the top secondary schools will be made available.

## Maritime Collision Detection and Avoidance

One of the WRSC competition challenges is to detect and avoid obstacles. Currently this is done through a gimbal-mounted front-facing camera.

We are looking for a student to investigate and evaluate a range of technology that could be used to solve this problem and use this knowledge to build a working obstacle detection system that could be mounted on next year's boat.

## AI Saling Boat Using ROS

The current boat control systems are written in Python and use simple deterministic logic and, as we add new capabilities the system gets more intertwined. This project would be to build a new, modular boat control system using the Robot Operating System (ROS). This would require the creation of ROS leaf nodes to handle sensor data the development of a ROS Computational Graph. You would also be encouraged to explore the powerful ROS logging tools.

## School Autonomous Boat Challenge

The IMechE and IET have made it a priority to encourage more students to become autonomous vehicle engineers and are keen to get school children interested in this. We already have an autonomous boat that was used to compete in 2019.

We are looking for a student to re-engineer the existing boat into a teaching tool that will introduce school children to the concepts around designing autonomous boats. In order to make this as accessible as possible the system could be based on the Graupner Micro Magic RC boat and powered using the BBC Micro-Bit. Whilst the development of the vehicle is important the project should also involve the development of a set of teaching resources, these can be tested on the visiting students during the open days as well as in local secondary schools. A mentor from one of the top secondary schools will be made available.

An important outcome is the development of a kit list and build instructions to allow schools to get involved. You should also consider where these could be raced (given the small size perhaps in a swimming pool?)

## EEC HVAC Integration

The EEC building is controlled via a [BACNet](https://en.wikipedia.org/wiki/BACnet) system which required expensive hard-wired controllers whilst home automation makes use of low cost wireless sensors communicating via [MQTT](http://mqtt.org).

This project, sponsored by the University, is to develop a gateway to enable low-cost (Arduino-based) sensors to control parts of the building. If successful this will be rolled out to both buildings and will allow students to build projects that integrate with these.

This work has already been started by [Oliver Bell](mailto:bello@coventry.ac.uk) who has open sourced his [BACprop MQTT to BACnet gateway](https://github.com/freshollie/bacprop) and is happy to answer any questions.

## Mobile Vehicle Detection and Fault reporting

Accurate tracking of traffic is becoming essential for the development of congestion detection systems. This project is to develop a low-cost mobile vehicle tracking tool.

The project will (probably) be based around either a Raspberry Pi 4 or Jetson Nano and will require the development and evaluation of a simple [ANPR](https://en.wikipedia.org/wiki/Automatic_number-plate_recognition) solution which can then be used to identify vehicles. Once this has been build and evaluated this could be used to build a system to detect issues such as:

1. Faulty vehicle lights.
2. Drivers on their smartphones.

## Smart Lock

Develop a smart lock for use on cases and bags.