Systems and methods for controlling the operation of a vehicle

1. A system for controlling operation of a vehicle, the system comprising:

one or more processors; and

a memory communicably coupled to the one or more processors and storing:

a trajectory-prediction module including instructions that when executed by the one or more processors cause the one or more processors to:

generate predicted trajectories of the vehicle using first trajectory predictors based, at least in part, on first inputs including at least past trajectory information for the vehicle and sensor data, when the vehicle is operating in one of a manual driving mode and a semi-autonomous driving mode;

generate predicted trajectories of a road agent that is external to the vehicle using second trajectory predictors based, at least in part, on second inputs including at least past trajectory information for the road agent and the sensor data; and

integrate the predicted trajectories of the road agent into the first inputs to iteratively update the predicted trajectories of the vehicle and integrate the predicted trajectories of the vehicle into the second inputs to iteratively update the predicted trajectories of the road agent; and

a control module including instructions that when executed by the one or more processors cause the one or more processors to control operation of the vehicle based, at least in part, on at least one of the iteratively updated predicted trajectories of the vehicle and the iteratively updated predicted trajectories of the road agent.

2. The system of claim 1 , wherein the trajectory-prediction module includes further instructions to integrate predicted trajectories of at least one other road agent external to the vehicle into the first inputs to iteratively update the predicted trajectories of the vehicle and to integrate the predicted trajectories of the at least one other road agent into the second inputs to iteratively update the predicted trajectories of the road agent.

3. The system of claim 1 , wherein the trajectory-prediction module includes further instructions to:

compute a first set of confidence scores for the respective predicted trajectories of the vehicle as a function of time within a first predetermined temporal horizon using a first confidence estimator that includes a first deep neural network and iteratively update the first set of confidence scores as the predicted trajectories of the vehicle are updated; and

compute a second set of confidence scores for the respective predicted trajectories of the road agent as a function of time within a second predetermined temporal horizon using a second confidence estimator that includes a second deep neural network and iteratively update the second set of confidence scores as the predicted trajectories of the road agent are updated.

4. The system of claim 3 , wherein one of the first trajectory predictors is a first probabilistic variational trajectory predictor that includes a third deep neural network and one of the second trajectory predictors is a second probabilistic variational trajectory predictor that includes a fourth deep neural network.

5. The system of claim 4 , further comprising:

a model-training module including instructions that when executed by the one or more processors cause the one or more processors to train at least one of the first, second, third, and fourth deep neural networks.

6. The system of claim 3 , wherein the trajectory-prediction module includes further instructions to compute the first and second sets of confidence scores based, at least in part, on a number of iterations that occur between the first and second trajectory predictors while the predicted trajectories of the vehicle and the predicted trajectories of the road agent are being iteratively updated.

7. The system of claim 1 , wherein the trajectory-prediction module includes further instructions to substitute, for the predicted trajectories of the vehicle, in the second inputs a deterministic trajectory of the vehicle to iteratively update the predicted trajectories of the road agent, when the vehicle is operating in an autonomous driving mode.

8. The system of claim 1 , wherein the control module includes instructions to control operation of the vehicle by planning a trajectory for the vehicle.

9. The system of claim 1 , wherein the iteratively updated predicted trajectories of the vehicle and the iteratively updated predicted trajectories of the road agent include at least one of specific predicted trajectories sampled from respective first and second probability distributions and parameters defining the respective first and second probability distributions.

10. The system of claim 1 , wherein the sensor data includes at least one of image data, Controller Area Network (CAN bus) data, Inertial Measurement Unit (IMU) data, light detection and ranging (LIDAR) data, radar data, and sonar data and the sensor data originates from at least one of the vehicle and the road agent.

11. A non-transitory computer-readable medium for controlling operation of a vehicle and storing instructions that when executed by one or more processors cause the one or more processors to:

generate predicted trajectories of the vehicle using first trajectory predictors based, at least in part, on first inputs including at least past trajectory information for the vehicle and sensor data when the vehicle is operating in one of a manual driving mode and a semi-autonomous driving mode;

generate predicted trajectories of a road agent that is external to the vehicle using second trajectory predictors based, at least in part, on second inputs including at least past trajectory information for the road agent and the sensor data;

integrate the predicted trajectories of the road agent into the first inputs to iteratively update the predicted trajectories of the vehicle and integrate the predicted trajectories of the vehicle into the second inputs to iteratively update the predicted trajectories of the road agent; and

control operation of the vehicle based, at least in part, on at least one of the iteratively updated predicted trajectories of the vehicle and the iteratively updated predicted trajectories of the road agent.

12. The non-transitory computer-readable medium of claim 11 , wherein the instructions to integrate the predicted trajectories of the vehicle into the second inputs to iteratively update the predicted trajectories of the road agent include further instructions to substitute, for the predicted trajectories of the vehicle, in the second inputs a deterministic trajectory of the vehicle to iteratively update the predicted trajectories of the road agent, when the vehicle is operating in an autonomous driving mode.

13. A method of controlling a vehicle, the method comprising:

generating predicted trajectories of the vehicle using first trajectory predictors based, at least in part, on first inputs including at least past trajectory information for the vehicle and sensor data when the vehicle is operating in one of a manual driving mode and a semi-autonomous driving mode;

generating predicted trajectories of a road agent that is external to the vehicle using second trajectory predictors based, at least in part, on second inputs including at least past trajectory information for the road agent and the sensor data;

integrating the predicted trajectories of the road agent into the first inputs to iteratively update the predicted trajectories of the vehicle and integrating the predicted trajectories of the vehicle into the second inputs to iteratively update the predicted trajectories of the road agent; and

automatically controlling operation of the vehicle based, at least in part, on at least one of the iteratively updated predicted trajectories of the vehicle and the iteratively updated predicted trajectories of the road agent.

14. The method of claim 13 , further comprising:

integrating predicted trajectories of at least one other road agent external to the vehicle into the first inputs to iteratively update the predicted trajectories of the vehicle; and

integrating the predicted trajectories of the at least one other road agent into the second inputs to iteratively update the predicted trajectories of the road agent.

15. The method of claim 13 , further comprising:

computing a first set of confidence scores for the respective predicted trajectories of the vehicle as a function of time within a first predetermined temporal horizon using a first confidence estimator that includes a first deep neural network and iteratively updating the first set of confidence scores as the predicted trajectories of the vehicle are updated; and

computing a second set of confidence scores for the respective predicted trajectories of the road agent as a function of time within a second predetermined temporal horizon using a second confidence estimator that includes a second deep neural network and iteratively updating the second set of confidence scores as the predicted trajectories of the road agent are updated.

16. The method of claim 15 , wherein one of the first trajectory predictors is a first probabilistic variational trajectory predictor that includes a third deep neural network and one of the second trajectory predictors is a second probabilistic variational trajectory predictor that includes a fourth deep neural network.

17. The method of claim 16 , further comprising:

training at least one of the first, second, third, and fourth deep neural networks.

18. The method of claim 15 , wherein computing the first and second sets of confidence scores is based, at least in part, on a number of iterations that occur between the first and second trajectory predictors while the predicted trajectories of the vehicle and the predicted trajectories of the road agent are being iteratively updated.

19. The method of claim 13 , further comprising:

substituting, for the predicted trajectories of the vehicle, in the second inputs a deterministic trajectory of the vehicle to iteratively update the predicted trajectories of the road agent, when the vehicle is operating in an autonomous driving mode.

20. The method of claim 13 , wherein controlling operation of the vehicle includes planning a trajectory for the vehicle.