Radar is the unglamorous sensor that the sensor-rich camp refuses to drop, and this claim explains why. UATC's grant US12013457B2 ("Systems and methods for integrating radar data for improved object detection in autonomous vehicles," issued June 18, 2024; inventors Raquel Urtasun, Bin Yang, Ming Liang, Sergio Casas, Runsheng Benson Guo) fences integrating radar into a learned detector alongside LiDAR — the explicit code G01S 13/865 is "systems combining radar and LiDAR."

The mechanism fuses radar's distinctive strengths into the detection pipeline: G01S 13/589 marks radar velocity measurement, and radar's long range and ability to penetrate rain, fog, and dust cover exactly the conditions where LiDAR and cameras degrade. A learned detector (G06N 3/045) that ingests radar gets robustness in bad weather and direct velocity that LiDAR has to infer. The claim fences that robustness-by-redundancy stance.

For the control-and-perception beat, this is a Camp-Lines data point on the maximalist side: keep every modality, fuse them, let the network exploit each one's complementary failure modes. It's the opposite of the vision-only thesis. And it's the same Urtasun-group lineage as UATC's earlier detection family — a consistent multi-sensor philosophy fenced across years.

From a portfolio angle, radar-inclusive fusion grants matter more as the field re-litigates the sensor question. Some players quietly added radar back after early dismissals; a 2024 grant explicitly fencing radar-LiDAR fusion stakes out that pragmatic position and is the kind of asset Aurora (which absorbed UATC's autonomy IP) would value for trucking, where bad-weather robustness is non-negotiable.

Caveats. Radar-LiDAR fusion has prior art in both automotive and defense sensing; the grant turns on the specific integration-and-detection method in claim 1, not on the idea of adding radar. The breadth depends on the recited fusion step. Read the independent claim for how radar enters the detector.

For the file: a radar-inclusive fusion grant marking the maximalist sensor camp. Pull US12013457B2 on PatentBear, read claim 1 for the radar-integration step, and file G01S 13/865 as the marker for radar-LiDAR fusion IP.