Accelerated Signal Processing of Burst-Mode Streamline Data for Low-Power Embedded Multi-Channel LiDAR Systems

Abstract

Autonomous driving vehicle such as car or auto-mated guided vehicle (AGV), requires sensors to detect obstacles in its surrounding environment and feed the information back to the vehicle control system. Light detection and ranging (LiDAR) sensors, which have attracted attention as effective sensors for autonomous driving, provide distance values measured within field of view using a laser to generate three-dimensional (3D) coordinates. 3D coordinates have the advantage of providing more precise measurements of an object's shape than other sensors and can also indicate the distance to the object. However, as the amount of data generated for a 3D point cloud and the number of sensors increases, the total amount of data to be processed increases. Therefore, the processing of this raw data in the main processor of an autonomous vehicle becomes a huge burden. In order to reduce the burden on the main processor of the autonomous vehicle, this paper intends to develop a real-time embedded system that primarily processes data from LiDAR sensors.