Active learning and validation system for vehicles

A system, comprising:

• one or more non-transitory computer-readable storage media including instructions;
• and
• one or more processors coupled to the non-transitory computer-readable storage media, the one or more processors configured to execute the instructions to:
• generate, using a path planner, a trajectory parameter about a trajectory associated with controlling a vehicle through a scenario, the trajectory parameter indicating an aspect about controlling the vehicle over the trajectory;
• compare the trajectory parameter of the trajectory against a validation parameter associated with a validation dataset to determine a level of similarity, wherein the validation parameter is based on human-based vehicle driving trajectory data, wherein the trajectory parameter additionally indicates quantitative measures of relationships between the vehicle and objects in an environment;
• and responsive to determining that the level of similarity indicates that the trajectory parameter is not similar to the validation parameter, provide training data associated with the scenario to a training dataset so that a subsequent parameter of a subsequent trajectory generated by the path planner and associated with the scenario satisfies the level of similarity against the validation parameter and flag the trajectory as being insufficiently similar to the human-based vehicle driving trajectory data,
• wherein the level of similarity is based on one or more predetermined criteria, and the predetermined criteria correspond to one or more driving parameters, the predetermined criteria including at least a legality score that whether indicates the trajectory complies with local laws.

A non-transitory computer-readable medium comprising instructions that, when executed by one or more processors of a computing device, cause the one or more processors to:

• generate, using a path planner, a trajectory parameter about a trajectory associated with controlling a vehicle through a scenario, the trajectory parameter indicating an aspect about controlling the vehicle over the trajectory;
• compare the trajectory parameter of the trajectory against a validation parameter associated with a validation dataset to determine a level of similarity, wherein the validation parameter is based on human-based vehicle driving trajectory data, wherein the trajectory parameter additionally indicates quantitative measures of relationships between the vehicle and objects in an environment;
• and
• responsive to determining that the level of similarity indicates that the trajectory parameter is not similar to the validation parameter, provide training data associated with the scenario to a training dataset so that a subsequent parameter of a subsequent trajectory generated by the path planner and associated with the scenario satisfies the level of similarity against the validation parameter and flag the trajectory as being insufficiently similar to the human-based vehicle driving trajectory data,
• wherein the level of similarity is based on one or more predetermined criteria, and the predetermined criteria correspond to one or more driving parameters, the predetermined criteria including at least a legality score that indicates whether the trajectory complies with local laws.