Imitation learning for trajectory planning
Trajectory planning include identifying state data based on sensor data from sensors of a vehicle. The state data are input to a machine-learning model to obtain parameters of a trajectory planner. The parameters are input to the trajectory planner to obtain a short term speed plan. The vehicle is autonomously controlled according to at least a portion of the short term speed plan.
1. A method for trajectory planning, comprising:
identifying state data based on sensor data from sensors of a vehicle;
inputting the state data to a machine-learning model to obtain parameters of a trajectory planner, wherein the parameters of the trajectory planner comprise at least one of a target timing headway parameter or a target radial distance, and wherein the machine-learning model is trained using snippets of driving data from different vehicles to imitate driving behavior, such that output parameters are configured to adapt trajectory planning based on driving context;
inputting the parameters to the trajectory planner to obtain a short term speed plan, wherein the short term speed plan comprises a sequence of positions and speeds for the vehicle based on the target timing headway parameter or the target radial distance; and
autonomously controlling the vehicle according to at least a portion of the short term speed plan.
2. The method of claim 1 , wherein the machine-learning model is trained using data samples extracted from the state data, and wherein each data sample constitutes a snippet having a predefined length.
3. The method of claim 1 , wherein the state data comprises data related to at least two of localization data, data related to autonomous driving, or data related to perceived other road objects.
4. The method of claim 1 , wherein the machine-learning model is trained using state data that meet a selection criterion.
5. The method of claim 4 , further comprising:
receiving a selection of the machine-learning model.
6. A device for trajectory planning, comprising:
a processor; and
a memory, the processor configured to execute instructions stored in the memory to:
identify state data based on sensor data from sensors of a vehicle;
input the state data to a machine-learning model to obtain parameters of a trajectory planner, wherein the parameters of the trajectory planner comprise at least one of a target timing headway parameter or a target radial distance, and wherein the machine-learning model is trained using snippets of driving data from different vehicles to imitate driving behavior, such that output parameters are configured to adapt trajectory planning based on driving context;
input the parameters to the trajectory planner to obtain a short term speed plan, wherein the short term speed plan comprises a sequence of positions and speeds for the vehicle based on the target timing headway parameter or the target radial distance; and
autonomously control the vehicle according to at least a portion of the short term speed plan.
7. The device of claim 6 , wherein the machine-learning model is trained using data samples extracted from the state data, and wherein each data sample constitutes a snippet having a predefined length.
8. The device of claim 6 , wherein the state data comprises data related to at least two of localization data, data related to autonomous driving, or data related to perceived other road objects.
9. The device of claim 6 , wherein the machine-learning model is trained using state data that meet a selection criterion.
10. The device of claim 6 , wherein the processor is further configured to execute instructions stored in the memory to:
receive a selection of the machine-learning model.
11. A non-transitory computer-readable storage medium, comprising executable instructions that, when executed by a processor, cause performance of operations, comprising:
identifying state data based on sensor data from sensors of a vehicle;
inputting the state data to a machine-learning model to obtain parameters of a trajectory planner, wherein the parameters of the trajectory planner comprise at least one of a target timing headway parameter or a target radial distance, and wherein the machine-learning model is trained using snippets of driving data from different vehicles to imitate driving behavior, such that output parameters are configured to adapt trajectory planning based on driving context;
inputting the parameters to the trajectory planner to obtain a short term speed plan, wherein the short term speed plan comprises a sequence of positions and speeds for the vehicle based on the target timing headway parameter or the target radial distance; and
autonomously controlling the vehicle according to at least a portion of the short term speed plan.
12. The non-transitory computer-readable storage medium of claim 11 , wherein the machine-learning model is trained using data samples extracted from the state data, and wherein each data sample constitutes a snippet having a predefined length.
13. The non-transitory computer-readable storage medium of claim 11 , wherein the state data comprises data related to at least two of localization data, data related to autonomous driving, or data related to perceived other road objects.
14. The non-transitory computer-readable storage medium of claim 11 , wherein the machine-learning model is trained using state data that meet a selection criterion.