ABSTRACT

 Autonomous vehicles are becoming a trend in recent years with advanced technology. In this project, we simulated a self-driving car system with Raspberry Pi 3 as the core processor. Besides essential lane following, we also constructed a system stimulating the system of UBER. A sky-view webcam locates the position of our car and target like what a GPS system does, and then PC calculates the route and give instructions to the car through wireless low-power XBEE S2C modules. Moreover, traffic signs near the lane can be processed and predicted by fine-tuned MobileNet, which tells the car either to go certain route or alter its speed. Integration requires overall understanding and cooperation of all team members, and error handling is needed as well. After fine-tuning and revisions, our system is now characterized by its exceptional stability and reliability.