The Secret to China Speed Is Simple (But Not Easy): I Help Fix Your Automotive Software Operations—Happy Developers, Happy Quality Managers, Hassle-Free ASPICE Compliance
This radar sensor is crazy 🤯 Have you ever seen 4D radar data? Now you have. 4D radar offers insane resolution. The „pointcloud“ remembers of sth lidar like. But radar data is very different. The angular resolution is limited compared to lidar. Do you know how the vertical resolution revolution compares?
Still impressive,... and maybe good enought.. all what is matter is cost and fact that OEM's are used to radar. However it offer similar capability to LiDAR with lower angular resolution and distance accuracies. Also easier to put into the car as plactics elements do not influence its fields of view. Eastetic argument.
While radars have their own advantages, their resolution is still insufficient for some basic ADAS functions. For example, accurate hazard height estimation is crucial for determining drivable areas and ensuring comfortable maneuvering. The cost reduction of FMCW Lidars with sufficient weather robustness should position them as the true redundant sensor for cameras.
Co-Founder of PERCIV AI - developing AI for radars | Guest researcher at TU Delft
7moI have seen and used multiple kinds of 4d radar data. I can confidently say that next gen radars, especially when paired with proper perception designed for them (khm Perciv AI khm) will put the low to mid category of the lidar lineup into a difficult position. Altos Radar deserves a shout out here for example