Systems and methods for training probabilistic object motion prediction models using non-differentiable prior knowledge
The present disclosure provides systems and methods for training probabilistic object motion prediction models using non-differentiable representations of prior knowledge. As one example, object motion prediction models can be used by autonomous vehicles to probabilistically predict the future location(s) of observed objects (e.g., other vehicles, bicyclists, pedestrians, etc.). For example, such models can output a probability distribution that provides a distribution of probabilities for the future location(s) of each object at one or more future times. Aspects of the present disclosure enable these models to be trained using non-differentiable prior knowledge about motion of objects within the autonomous vehicle's environment such as, for example, prior knowledge about lane or road geometry or topology and/or traffic information such as current traffic control states (e.g., traffic light status).
1. A computer-implemented method for autonomous vehicle motion control, the method comprising:
obtaining sensor data for an environment comprising an object;
processing the sensor data with a machine-learned object motion prediction model to obtain a predicted location probability distribution for a future location of the object at one or more future times, wherein the machine-learned object motion prediction model has been trained by performing a REINFORCE gradient estimation technique to determine an approximate gradient of an expected loss that is a function of a non-differentiable prior knowledge reward function that encodes prior knowledge about motion of the object;
generating a vehicle trajectory based, at least in part, on the predicted location probability distribution for the future location of the object at the one or more future times; and
controlling motion of an autonomous vehicle based on the generated vehicle trajectory.
2. The computer-implemented method of claim 1 , wherein:
the sensor data for the environment comprises real world sensor data collected by sensors associated with the autonomous vehicle while the autonomous vehicle is operating; and
the non-differentiable prior knowledge reward function comprises a sample trajectory with ground truth data associated with the object.
3. The computer-implemented method of claim 1 , wherein the prior knowledge about motion of the object comprises prior knowledge about lane geometry, road topology, or traffic rules within the environment.
4. The computer-implemented method of claim 1 , wherein generating the vehicle trajectory based, at least in part, on the predicted location probability distribution for the future location of the object at the one or more future times comprises:
generating a trajectory score for each of a plurality of sample trajectories based, at least in part, on the predicted location probability distribution for the future location of the object at the one or more future times; and
selecting a sample trajectory from the plurality of sample trajectories based on the trajectory scores.
5. The computer-implemented method of claim 4 , wherein generating the trajectory score for each of the plurality of sample trajectories based, at least in part, on the predicted location probability distribution for the future location of the object at the one or more future times comprises:
determining a reachable area for the object, wherein the reachable area is defined by a set of one or more reachable lanes that are reachable from a current location of the object while observing traffic rules; and
generating the trajectory score for each of the plurality of sample trajectories based, at least in part, on the reachable area.
6. The computer-implemented method of claim 5 , wherein:
the generated trajectory score is increased when the sample trajectory stays within the reachable area; and
the generated trajectory score is decreased when the sample trajectory exits the reachable area.
7. The computer-implemented method of claim 5 , wherein:
the generated trajectory score is increased when the sample trajectory stays within the reachable area and a ground truth trajectory associated with the object stays within a ground truth reachable area; and
the generated trajectory score is decreased when the sample trajectory exits the reachable area and the ground truth trajectory associated with the object stays within the ground truth reachable area.
8. The computer-implemented method of claim 5 , wherein the traffic rules comprise lane infraction rules that prohibit crossing a solid line or entering a lane having an opposite traffic flow direction.
9. The computer-implemented method of claim 5 , wherein the traffic rules comprise adhering to a current traffic control state provided by a traffic control device.
10. The computer-implemented method of claim 4 , wherein generating the trajectory score for each of the plurality of sample trajectories based, at least in part, on the predicted location probability distribution for the future location of the object at the one or more future times comprises:
determining, for each sample trajectory, whether the sample trajectory intersects with a route associated with an ego vehicle, wherein the trajectory score is generated based on whether the sample trajectory intersects with the route.
11. The computer-implemented method of claim 10 , wherein the route associated with the ego vehicle comprises a current motion plan for the ego vehicle.
12. The computer-implemented method of claim 1 , wherein the machine-learned object motion prediction model comprises one or more neural networks.
13. The computer-implemented method of claim 4 , wherein the one or more future times comprise a plurality of future times.
14. The computer-implemented method of claim 1 , wherein the machine-learned object motion prediction model comprises a spatially-aware graph neural network combined with a multi-layer perceptron parameterized as a mixture of Gaussians.
15. The computer-implemented method of claim 1 , wherein the training of the machine-learned object motion prediction model comprises:
modifying one or more values of one or more parameters of the machine-learned object motion prediction model based at least in part on the approximate gradient of the expected loss.
16. The computer-implemented method of claim 15 , wherein modifying the one or more values of the one or more parameters of the machine-learned object motion prediction model based at least in part on the approximate gradient of the expected loss comprises modifying the one or more values of the one or more parameters of the machine-learned object motion prediction model based at least in part on a heuristically chosen closest matching mode.
17. A computing system, comprising:
one or more processors; and
one or more non-transitory computer-readable mediums that collectively store instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising:
obtaining sensor data for an environment comprising an object;
processing the sensor data with a machine-learned object motion prediction model to obtain a predicted location probability distribution for a future location of the object at one or more future times, wherein the machine-learned object motion prediction model has been trained by performing a REINFORCE gradient estimation technique to determine an approximate gradient of an expected loss that is a function of a non-differentiable prior knowledge reward function that encodes prior knowledge about motion of the object;
generating a vehicle trajectory based, at least in part, on the predicted location probability distribution for the future location of the object at the one or more future times; and
controlling motion of an autonomous vehicle based on the generated vehicle trajectory.
18. The computing system of claim 17 , wherein:
the sensor data for the environment comprises real world sensor data collected by sensors associated with the autonomous vehicle while the autonomous vehicle is operating; and
the non-differentiable prior knowledge reward function compares a sample trajectory with ground truth data associated with the object.
19. The computing system of claim 17 , wherein the prior knowledge about motion of the object comprises prior knowledge about lane geometry, road topology, or traffic rules within the environment.
20. An autonomous vehicle computing system, comprising:
one or more processors; and
one or more non-transitory computer-readable mediums that collectively store:
a machine-learned object motion prediction model configured to process sensor data descriptive of an environment comprising an object to generate a predicted location probability distribution for a future location of the object at one or more future times, wherein the machine-learned object motion prediction model has been trained by performing a REINFORCE gradient estimation technique to determine an approximate gradient of an expected loss that is a function of a non-differentiable prior knowledge reward function that encodes prior knowledge about motion of the object; and
instructions that, when executed by the one or more processors, cause the computing system to perform operations, the operations comprising:
obtaining the sensor data descriptive of the environment comprising the object;
processing the sensor data with the machine-learned object motion prediction model to generate the predicted location probability distribution for the future location of the object at the one or more future times;
generating a vehicle trajectory based, at least in part, on the predicted location probability distribution for the future location of the object at the one or more future times; and
controlling motion of an autonomous vehicle based at least in part on the generated vehicle trajectory.