OASD: An Open Approach to Self-Driving Vehicle

Abstract

Building autonomous vehicles is an arduous task and requires a lot of technical understanding and cognizance. Grasping the essentials of such technologies and finding the perfect combination of hardware and software implementation is key to building autonomous vehicles. As we move towards autonomous vehicles, How do we find a solution to building an autonomous vehicle? To answer this question, we break it down into different hardware and software requirements. We present a hands-on solution for building autonomous vehicles by using an open-source platform Autoware which is based on ROS (Robot Operating System). Our initial implementation consists of detailing the overall system requirements, sensors calibration, point cloud map generation, waypoint generation and following, and finally vehicle control. We have also conducted the simulation test of localization and path following based on the point cloud map and waypoint generated. We believe our implementation provides a general understanding for building autonomous vehicles in the academic as well as in research field.