These two patents are the same financial question asked twice: can you get ADAS-grade perception without the most expensive sensor? On June 9, 2026, Korea Advanced Institute of Science and Technology was granted US12651445B2, "System and method for perceiving 3-D environment using camera and radar" (CPC G01S 13/867 sensor fusion, G06V 20/58 vehicle scene understanding). The same day, Neural Propulsion Systems was granted US12650505B2, "Multi-frequency radar array systems and sensor fusion for seeing around corners in autonomous driving" (CPC G01S 13/931 vehicle radar, G01S 13/865 radar-sensor fusion).
Why frame two sensor patents as a margin story? Because the sensor suite is one of the costliest parts of an ADAS or autonomy system, and radar is materially cheaper than LiDAR. Every method that squeezes more capability out of radar — fusing it with cameras for 3-D perception, or arraying multiple frequencies to see around occlusions — is a way to hit a capability target at a lower sensor bill of materials. That's a direct lever on ADAS unit cost.
The guide moved — quietly — on sensor strategy across the industry, and these patents are part of why. The expensive-LiDAR-everywhere assumption has been under pressure precisely because it wrecks the cost model for mainstream vehicles. Camera-radar fusion and smarter radar are the engineering response to a financial constraint: ADAS has to be affordable to ship at volume.
Scope discipline: KAIST's grant covers a specific camera-radar 3-D perception method; Neural Propulsion's covers a specific multi-frequency around-the-corner radar arrangement. Neither claims radar perception broadly. For an analyst, the value is as evidence of a sector-wide push toward cheaper sensor suites, held in this case by a university and a specialist — not by the automakers, which is itself a sourcing signal.
The financial confirmation is in the cost trend: as cheaper-sensor ADAS architectures mature, the per-vehicle sensor bill should fall relative to capability, supporting ADAS gross margin. That cost trajectory shows up indirectly in automaker and supplier cost-of-revenue, trackable through the SEC filing evidence index. These radar patents are the upstream attempt to make ADAS pencil out.