Enabling claim or aspiration? Nissan's grant US12649492B2, "Constraint-based speed profile" (issued June 9, 2026), lands on the enabling side, and the CPC trail proves it: B60W 60/0027 (autonomous operation in specific situations), B60W 30/0956 (predicting or estimating in traffic situations), and B60W 40/105 (vehicle speed). These are the codes for the part of the autonomy stack that turns a route into actual pedal-and-brake commands — the planning layer, not the marketing layer.
Here is the problem the patent solves. Planning a self-driving trajectory is not just choosing a path through space; it is choosing a speed at every point along that path. Go too fast into a curve and you violate comfort or lose traction; brake too early and you are the annoying robot car blocking traffic. The speed profile is the function that assigns a target speed to each point, and getting it right is one of the genuinely hard, genuinely shippable problems in autonomy.
"Constraint-based" is the mechanism, and it is a specific design choice. Rather than hand-tuning speeds or running a single optimizer with everything baked in, the method represents the requirements as explicit constraints — maximum speed from the limit, maximum lateral acceleration from the curvature, minimum following distance, comfort bounds on jerk — and then finds a speed profile that satisfies them. The advantage is that constraints are inspectable and composable: you can add, remove, or weight them without rewriting the planner, and you can prove which constraint is binding at any moment.
What is novel versus standard practice? Constrained trajectory optimization is well-trodden ground in robotics and AV research, so the defensible novelty is in the particular formulation this grant claims — which constraints, how they are combined, how the profile is generated. The independent claim defines that method; reading the abstract as "Nissan owns AV speed planning" would overstate it badly. The scope is the specific constraint-based generation technique, fenced by the dependent claims.
The assignee — Nissan North America — and inventors associated with AV motion planning point to where this sits: inside an automaker's own autonomy program, not a supplier's licensable module. That is the opposite signal from the tier-1 steering patent elsewhere on this site; a captive AV-planning patent is more likely to stay in-house than to define an industry standard.
For the B60W autonomy class, motion-planning patents like this are the substance under the hype. When a company says its car "drives itself," the question to ask is what the planning claims actually enable. Here, the answer is concrete: a method for assigning safe, comfortable speeds along a path by satisfying explicit constraints — an enabling capability, precisely scoped, and now granted.