Tom Lewis

Issie

Issie (Interactive Schematic Simulator with Integrated Editor) is an application for digital circuit design and simulation. It is targeted at students and hobbyists that want to get a grasp of Digital Electronics concepts in a simple and fun way. As part of the third year High Level Programming module, I worked as part of a group to improve Issie, which is written in F#. We focused on the drawblock in particular, which allows the user to place and move components, and connect them together. We extended the existing functionality to allow components to be rotated and flipped, as well as allowing the user to change the positions of ports on components. Another element of the task was to refactor the existing code, making use of discriminated union types where possible.

The code for this project can be found on GitHub.

Close Window

C Compiler

As part of the second year Instruction Architectures and Compilers module, I worked with a partner to create a compiler for a subset of the ANSI C language. The compiler uses Flex for lexing, Bison for parsing, and C++ for everything else. The target language is the MIPS ISA.

The code for this project can be found on GitHub.

Close Window

Machine Learning

Through various modules at university and through other side projects, I have used machine learning in a number of ways. Some of these include:

Decision trees - Using decision trees to work out which room in a building someone is in. The data used was the strength of different WiFi signals.

Neural networks - Creating a custom neural network library, and using PyTorch to predict house prices.

Image classification - Classifying different species of fish using CNNs, also implemented using PyTorch.

The code for some of these project can be found on GitHub.

Close Window

Embedded Systems

The third year Embedded Systems module was comprised of two significant courseworks (each completed by a group of four):

Internet of Things Device - For this coursework we created a work/study environment sensor called IrisAir. It tracks the ambient air temperature, humidity, light intensity, and colour spectrum and uses machine learning to notify the user when the current working conditions are suboptimal. It then advises how to improve the conditions.

Music Synthesiser - This coursework involved writing complex low-level code for a music synthesiser. The code interacts with the synthesiser hardware to check events such as key presses, button presses and knob turns. With all that information it then calculates the appropriate sound to make. The implementation uses a combination of threads and interrupts, and much care was taken to ensure there was no chance of deadlock, and all inter-task dependancies were appropriately managed. This was achieved through a combination of atomic accesses, semaphores, mutexes and critical sections.

The code for these projects can be found on GitHub.

Close Window

Wall-eee

My second year Computer Engineering Design Project module involved writing software for a Mars rover. The rover consisted of a camera, FPGA, Arduino and other electronic components. The rover was split in to 5 subsystems, and one person worked on each subsystem. I was responsible for the 'Command' subsystem, which involved creating a user interface to send commands to the rover and view its position and the positions of objects it has identified.

The code for this project can be found on GitHub.

Close Window

MIPS CPU

As part of the second year module Instruction Architectures and Compilers, I helped a team produce a working MIPS CPU in SystemVerilog, along with a comprehensive testbench. I was personally responsible for writing the SystemVerilog testbech, along with bash scripts and C tests.

The code for this project can be found on GitHub.

Close Window