A toy truck I took apart and programmed to chase anything that moved. Would have been easier with a few more sensors but was still a fun project. Used a genetic algorithm so it could learn the most efficient way to navigate. It really needs a wheel encoder to work more efficiently. Also experimented with averaging the mean and median of the white Gaussian noise to average the sensors values, but eventually chose to use the more system taxing invariant extended kalmen filter. Overkill I know, but I plan on building an unusual uav.