Two CPC codes on a system claim is a study in economy. Aurora Operations' grant US12358524B2 ("Perception system for an autonomous vehicle," issued July 15, 2025; inventors Nemanja Djuric, Shivam Gautam, Peter M. Kingston, Chi-Kuei Liu) classifies into exactly B60W 60/001 — autonomous driving control responsive to detected objects — and G06V 10/82 — neural-network image processing. Perception in, control out, learned in the middle. That's the whole claim's shape in two codes.

The mechanism is a learned perception system whose output drives the vehicle: G06V 10/82 marks the neural processing of sensor imagery, B60W 60/001 marks the autonomous control that consumes the perception. The tightness of the classification suggests the claim is a focused system claim — likely a specific architecture or data-flow for turning raw sensing into a control-ready environmental model — rather than a sprawling everything-claim.

“An autonomous platform can obtain sensor data descriptive of an actor in an environment of an autonomous vehicle and at least a portion of the environment of the autonomous vehicle that does not include the actor, the sensor data comprising at least one sweep of the environment of the autonomous veh…”— U.S. Patent No. 12,358,524 source

For the control-and-perception beat, a 2025 grant titled simply "Perception system for an autonomous vehicle" from Aurora is a marker of portfolio maturity: by this point the company is fencing whole-system perception claims, not just individual primitives like trailer tracking or coherent-LiDAR classification. The system-level title and the senior inventor list suggest a consolidating claim that ties Aurora's perception components together.

From a portfolio angle, this caps a multi-year Aurora perception lineage that runs through the Blackmore FMCW hardware (2022), the phase-coherent classification grant (2024), and the inherited UATC detection family. A system-level perception patent is the umbrella over those pieces — the kind of claim a maturing autonomy company files to fence the integrated whole once the parts are in place.

Caveats. A two-code classification is tight but not self-explaining; the breadth lives entirely in claim 1's recited system elements. AV perception is a crowded field, so the grant turns on the specific architecture, not on the existence of a perception system. Read the independent claim for which components and data-flows it makes mandatory.

For the file: a consolidating, system-level AV-perception grant capping Aurora's perception portfolio. Pull US12358524B2 in the patent record, read claim 1 for the recited system architecture, and read it as the umbrella over Aurora's earlier sensor and detection IP.