Hardware-accelerated Ethernet Interface for Automotive LiDAR Sensors

The automotive industry keeps pushing technology towards fully autonomous vehicles. In the near future, cars are expected to be sufficiently reliable, affordable, and common to simplify several driving tasks. To recreate the surrounding environment in real-time, an autonomous vehicle requires reliable
and accurate multi-sensor perception systems, which include RADARs, Cameras, and LiDAR sensors. LiDAR sensors are becoming very popular since they are able to create 2D/3D representations of the car’s vicinity in different light and weather conditions, even at high speeds. However, to achieve highresolution
visualizations, a LiDAR sensor must provide millions of data points to be processed, which often requires powerful processing systems. With this article, we propose an accelerated interface for LiDAR sensors composed of a data packet decoder and a reconstruction system, fully deployed on an embedded
reconfigurable hardware platform with field-programmable gate array (FPGA) technology. The proposed FPGA-based solution is able to interface different LiDAR sensors at the same time, outperforming software-only approaches by reaching lower latency and high performance results.