Neural network based vehicle dynamics model

A method comprising:

• training a machine learning system with a training dataset corresponding to a desired autonomous vehicle simulation environment;
• receiving vehicle control command data and vehicle status data;
• by use of the trained machine learning system, generating predicted vehicle acceleration data based on the vehicle control command data and the vehicle status data, wherein simulated vehicle dynamics data comprises the predicted vehicle acceleration data;
• providing the simulated vehicle dynamics data to an autonomous vehicle simulation system implementing the autonomous vehicle simulation environment;
• and
• conducting an iteration process periodically, wherein the iteration process comprises:
• receiving the vehicle status data modified by data produced by the autonomous vehicle simulation system;
• and
• by use of the trained machine learning system, generating modified simulated vehicle dynamics data based on the vehicle control command data and the modified vehicle status data.

A non-transitory machine-useable storage medium embodying instructions which, when executed by a machine, cause the machine to:

• train a machine learning system with a training dataset corresponding to a desired autonomous vehicle simulation environment;
• receive vehicle control command data and vehicle status data;
• by use of the trained machine learning system, generate predicted vehicle acceleration data based on the vehicle control command data and the vehicle status data, wherein simulated vehicle dynamics data comprises the predicted vehicle acceleration data;
• provide the simulated vehicle dynamics data to an autonomous vehicle simulation system implementing the autonomous vehicle simulation environment;
• and
• conduct an iteration process periodically, wherein the iteration process comprises:
• receiving the vehicle status data modified by data produced by the autonomous vehicle simulation system;
• and
• by use of the trained machine learning system, generating modified simulated vehicle dynamics data based on the vehicle control command data and the modified vehicle status data.