Tricycle Localization

Tyler Folsom, Computer & Software Systems/Electrical Engineering

The Elcano Project makes automated vehicle systems available to students and hobbyists as open-source. We focus on automating electric bicycles and tricycles at a price point that can make them the automated vehicle of choice. The localization system runs on an Arduino microcontroller and primarily uses GPS and dead reckoning. The two position estimates are combined by a fuzzy filter. Future work will add accelerometers and gyroscopes and a Kalman filter. We have some simple maps of UW Bothell and a few other areas. The vehicle is designed to find its position on the map. However, testing has been difficult. A team of EE capstone students have designed interface hardware and software that connects the Elcano microprocessors to the open-source CARLA driving simulator. The interface box lets us drive a virtual vehicle by feeding sensors and actuators with data from the simulator.