Controlling a sliding contact on an electric vehicle with computer vision and AI
Abstract: Emission from road vehicles is a massive problem and contributes to climate change on our planet. One solution that people are turning to is electrical propulsion instead of fossil fuel. There are, however, problems with putting big batteries on road vehicles. They are expensive to build, they require rare minerals, and the process of creating batteries emits plenty of greenhouse gas. To reduce the need of big batteries, Elonroad is creating a way of charging road vehicles while driving. This works by putting rails in roads and sliding contacts underneath vehicles. For this to work the sliding contacts and the rail needs to stay aligned while driving. In this thesis the problem is solved by controlling the sliding contacts position with use of a camera, machine learning and a controller. The proposed structure is to use a pre-trained neural network called MobileNet together with a custom neural network to estimate the position of the sliding contact. The estimated position is then used as input to a PID controller that controls the position of the sliding contact with a motor.
AT THIS PAGE YOU CAN DOWNLOAD THE WHOLE ESSAY. (follow the link to the next page)