Development

Here is a collection of projects of significance I have undertaken. Each project provides overviews of the key points, but more detail can be found in published work, as linked in the pages.

Ant World A grid based simulator in Python based on scans for a real ant testing site.
	3D printed tactile dataset Benchmarking tactile texture classification tasks in a replicatable way
Texture and friction classifiction A comparison between the two types of tactile sensors we have worked on
	Quadruped evolution Project for creating a CAD quadruped from scratch and evolving gait controllers
Tactile Biped Using the PressTip foot sensor, we were able to gather a dataset of poses and train a model to recognise its 3D orientation based on tactile readings
	PressTip multimodal sensor The construction of the PressTip sensor, an electrical multimodal tactile sensor.
TacTip construction The TacTip sensor construction and experimentation across various designs.
	GPS Ant robot This project worked on constructing a large chassis that could be used for outdoor exploration to test ant inspired algorithms. This project was part of a research assistant role
Bio-Inspired Navigation for Varied Terrain This project uses the Wheg chassis developed in previous projects, however investigates aspects of autonomous navigation behaviours for traversing complex terrain.
	Deep learning for autonomous navigation on a small robots This project was part of the University of Sussex Junior Research Associate scheme
Genetic algorithm robot walking optimization Genetic Algorithms use random mutations to within a Genotype which is assessed using a fitness function. In this case fitness being the Genotype which walks the best.
	Exploring planetary terrain In this project we explore different hardware options available for space exploration This started as a tracked rover controlled via a network.