Cameras-only is a bet; rich radar is the counter-bet, and GM Cruise's application US20230324539A1, "Multi-frequency micro-Doppler processing and fusion" (published October 12, 2023), is firmly in the radar camp. Micro-Doppler is the subtle frequency modulation that a target's moving parts — a pedestrian's swinging arms, a cyclist's spinning wheels — impose on a radar return. Fusing those signatures across multiple frequencies is the claimed method, and the CPC (G01S 13/865, 13/89, 13/931, with lidar G01S 17/931) is squarely radar-centric.
This is an enabling claim, not an aspirational one, because micro-Doppler classification is a real and specific capability radar can deliver that cameras struggle with — distinguishing a person from a pole by motion signature, even in poor light. Claiming the multi-frequency fusion of those signatures is claiming a concrete signal-processing technique. The sensor modality and the physical phenomenon are both named in the filing, which is exactly what an enabling AV radar claim should do.
The skeptic's discipline still applies. This is an A1 application, so the issued scope will be narrower, and micro-Doppler processing has prior art in defense and gesture-sensing domains that examiners will weigh. The defensible novelty is the specific multi-frequency fusion method, and the dependents that pin down how the frequencies combine and how the result feeds classification are where the protection will harden.
Strategically, GM Cruise filing this in October 2023 fits a builder hedging against the camera-only thesis by extracting more information from radar — a sensor that works in rain, fog, and darkness. The enabling-versus-aspirational verdict: enabling, because the method targets a real radar-distinct capability. Label it an application, watch the fusion-method dependents in prosecution, and read it as evidence that micro-Doppler classification had moved into serious AV IP by late 2023.