Control of vehicle automated driving operation with independent planning model and cognitive learning model

A system for controlling an automated driving operation in a vehicle, the system comprising:

• a controller operatively connected to the vehicle, the controller having a processor and tangible, non-transitory memory on which instructions are recorded;
• a computing unit in communication with the controller and having respective independent models for executing planning and learning in the automated driving operation, including a vehicle planning model and a cognitive learning model;
• wherein the computing unit includes a semantic layer adapted to act as a bridge between the cognitive learning model and the vehicle planning model, the semantic layer including a first data adaptor and a second data adaptor;
• wherein the first data adaptor is configured to transform real traffic data to a respective equivalent abstract representation such that it can be used by the cognitive learning model to generate a humanized reward model;
• wherein the vehicle planning model is adapted to determine a trajectory plan based in part on the humanized reward model;
• and
• wherein the controller is adapted to execute the automated driving operation based in part on the trajectory plan.

A method for controlling an automated driving operation of a vehicle having a controller with a processor and tangible, non-transitory memory, the method comprising:

• setting up respective independent models for executing planning and learning in the automated driving operation, including a vehicle planning model and a cognitive learning model;
• generating a semantic layer to act as a bridge between the cognitive learning model and the vehicle planning model, the semantic layer including a first data adaptor, a second data adaptor and a third data adaptor;
• obtaining real traffic data, the semantic layer employing a Markov Decision Process model;
• transforming the real traffic data to a respective equivalent abstract representation such that it can be used by the cognitive learning model to generate a humanized reward model, via the first data adaptor;
• determining a trajectory plan, via the vehicle planning model, based in part on the humanized reward model;
• obtaining pairs of state and action variables for the vehicle, via the vehicle planning model;
• transforming the pairs of state and action variables to the respective equivalent abstract representation in the semantic layer, via the second data adaptor, the respective abstract representation being transmitted to the humanized reward model;
• transmitting respective data from the humanized reward model to the vehicle planning model, via the third data adaptor;
• and
• executing, via the controller, the automated driving operation based in part on the trajectory plan.