A scanning LiDAR, by its inherent nature, generates a great deal of raw digital data. To generate 3D imagery in real time, the data must be processed as quickly as possible. One method of discerning time-of-flight of a laser pulse for a LiDAR application is correlating a Gaussian pulse with a discretely sampled waveform from the LiDAR receiver.
For the Goddard Reconfigurable Solid-state Scanning LiDAR (GRSSLi), an FPGA module was developed to process an arbitrary number of waveforms quickly and in parallel, enabling a high-resolution, 200-KHz time-of-flight solution. For GRSSLi, up to eight crosscorrelation engines were instantiated within an FPGA to process the discretely sampled transmit and receive pulses from the LiDAR receiver and measure the time-of-flight of a laser pulse at 20-picosecond resolution. The number of engines is only limited by the resources within the FPGA fabric, and is configurable with a constant. Thus, the timeof- flight measurement rates could go well beyond the 200-KHz requirement from GRSSLi.
The engines are efficiently designed to use the least amount of FPGA resources possible. This innovation allows a system to process an almost limitless number of received laser pulses for LiDAR applications in real time, limited only by the available FPGA resources.
This work was done by Nathaniel Gill of Goddard Space Flight Center and Mark Giza of the U.S. Army Research Laboratory.