Systems and methods for vehicle motion planning based on uncertainty

A computing system, comprising:

• one or more processors;
• and
• one or more memories including instructions that, when executed by the one or more processors, cause the one or more processors to perform operations, the operations comprising:
• obtaining scene data descriptive of one or more objects within a surrounding environment of an autonomous vehicle;
• determining a plurality of subproblems from the scene data, wherein a subproblem of the plurality of subproblems corresponds to a potential interaction of the autonomous vehicle with an object of the one or more objects within the surrounding environment of the autonomous vehicle, wherein the subproblem is associated with a precomputed solution comprising a function that governs the potential interaction;
• identifying a travel decision based at least in part on the scene data and plurality of subproblems, wherein the travel decision is indicative of a discrete vehicle action with respect to the surrounding environment of the autonomous vehicle;
• generating a branching policy based on the subproblem and the travel decision, wherein the branching policy comprises a plurality of nodes, wherein a node of the plurality of nodes is indicative of the potential interaction with the object within the surrounding environment of the autonomous vehicle;
• determining a cost for the branching policy using the precomputed solution associated with the subproblem;
• selecting the travel decision based on the cost for the branching policy at least in part;
• and
• initiating a motion of the autonomous vehicle based at least in part on the travel decision.

An autonomous vehicle comprising:

• one or more processors;
• and
• one or more tangible, non-transitory, computer readable media that store instructions that when executed by the one or more processors cause the one or more processors to perform operations comprising:
• obtaining scene data descriptive of one or more objects within a surrounding environment of the autonomous vehicle;
• determining a plurality of subproblems based at least in part on the scene data, wherein a subproblem of the plurality of subproblems corresponds to a potential interaction of the autonomous vehicle with an object of the one or more objects within the surrounding environment of the autonomous vehicle, wherein the subproblem is associated with a precomputed solution comprising a function that governs the potential interaction;
• generating a branching policy based at least in part on the plurality of subproblems, wherein the branching policy comprises a plurality of nodes, and wherein a node of the plurality of nodes corresponds to a travel decision and is indicative of the potential interaction with the object in the surrounding environment of the autonomous vehicle;
• determining a cost for the branching policy using the precomputed solution associated with the subproblem;
• selecting the travel decision based at least in part on the cost for the branching policy;
• and
• initiating a motion of the autonomous vehicle based at least in part on the travel decision.