Radar vs. LiDAR: Why Radar Wins for Autonomous Driving #shorts

Added: 2026-05-19

[00:00:00]How would lidar compare with radar in this application? Radar would be significantly a better sensor for this application because it's not only the sensor that matters, it's how much data you're generating and the computation that is required to process that data.

[00:00:15]The beauty of 4D imaging radar is that every point in the point cloud has a Doppler information and we could easily do static dynamic segmentation. We could turn off all the static points in the environment, which is essentially most of your points and only focus on the dynamic points. So we could focus on things that are moving and the computation computational complexity of doing it with the radar is orders of magnitude less than doing it with the lidar. So lidar is more expensive than the radar, but that's not the whole story. You have to have even a more expensive processor to process the point cloud coming from the radar from the lidar. But using the radar, you could do the processing at the edge. So Like inside our sensor?

[00:01:03]>> Inside the sensor, exactly. So our Z pulse platform, for example, or our Z prime is very capable. It could do quite a bit of processing at the edge cuz we have a very powerful FPGA. We could send out an object list uh which could include you know, the track number, the number of vehicles, the count, the duration, the speed, the location, >> Which where they are, the lane, everything. Exactly. type of object as well. classification all of that. Yeah, that could come out as an object list.

[00:01:37]And that's a very low data rate, very low data in the kilobit per second type of range. You don't need like massive point cloud coming out.