Vehicle Type Specific Waypoint Generation
This addresses a limitation in downstream planning applications for autonomous driving by improving the utility of behavior models that lack vehicle information.
The paper tackled the problem of generating vehicle-type specific waypoint sequences by conditionally specializing a generic probabilistic behavior model using vehicle-specific value functions, resulting in more physically plausible sequences compared to vehicle-agnostic methods.
We develop a generic mechanism for generating vehicle-type specific sequences of waypoints from a probabilistic foundation model of driving behavior. Many foundation behavior models are trained on data that does not include vehicle information, which limits their utility in downstream applications such as planning. Our novel methodology conditionally specializes such a behavior predictive model to a vehicle-type by utilizing byproducts of the reinforcement learning algorithms used to produce vehicle specific controllers. We show how to compose a vehicle specific value function estimate with a generic probabilistic behavior model to generate vehicle-type specific waypoint sequences that are more likely to be physically plausible then their vehicle-agnostic counterparts.