Self-generated 3D point cloud with raw data from FMCW radar

Point clouds are used everywhere, and they have become vital in numerous applications such as robotics, autonomous vehicles, and virtual reality. It is possible to generate a point cloud with a Frequency Modulated Continuous Wave (FMCW) radar. There are two ways: On-chip data of the point cloud from the radar or transforming the raw data of the radar into a point cloud externally of the radar. In this paper, I will focus on the latter part by recording the raw data with an FMCW radar and applying pre-processing steps on the raw data to find the points. To validate my implementation, my generated point cloud will be compared to the Robot Operating System (ROS) provided by Texas Instrument, which shows the point cloud by receiving on-chip data from the radar. The goal of this paper is to research the possibilities of improving the hardware limitations of the radar by doing computations outside of the radar.