Velocity is a roadmap tell, and the 2024 trajectory-planning record points squarely at the robotaxi builders. Across PatentBear's results for the year, Zoox and Waymo led the named-assignee count by a wide margin, with Motional, UATC/Aurora, NVIDIA, and GM Cruise forming the next tier. When two companies dominate filings in the planning class, that concentration says the planning layer — deciding the path, not perceiving the scene — was a priority investment area for the urban-autonomy players.
The class composition is consistent and telling. The cluster sits in B60W 60/001, 60/0011 and 60/0015 (autonomous-behavior planning) braided with G06N 20/00 and G06N 3/08 (machine learning), marking a year in which learned and prediction-based planning had largely displaced purely hand-engineered planners. A representative example is application US20240288868A1 on data-driven prediction-based trajectory planning, which captures the prediction-then-plan architecture the cluster favors.
For portfolio mapping, the interpretation is that 2024 was a planning-and-prediction year for the robotaxi cohort. Perception IP continued, but the dense new activity moved downstream to the question of what to do with the perceived scene — how to predict other agents and plan a safe trajectory among them. That is the layer where urban autonomy is hardest, so the concentration there tracks the genuine engineering frontier rather than fashion.
The analyst's caveats hold: a large share of 2024 planning results carry no assignee in the index, understating totals, and filing volume is directional rather than a quality measure. But the Zoox-and-Waymo lead is robust and dated cleanly to 2024. Read the cluster as the tell it is — the named-assignee concentration in B60W 60/00 planning classes marks exactly which builders were pouring effort into the decision layer of self-driving.