Multi-perspective system and method for behavioral policy selection by an autonomous agent
A system and a method for autonomous decisioning and operation by an autonomous agent includes: collecting decisioning data including: collecting a first stream of data includes observation data obtained by onboard sensors of the autonomous agent, wherein each of the onboard sensors is physically arranged on the autonomous agent; collecting a second stream of data includes observation data obtained by offboard infrastructure devices, the offboard infrastructure devices being arranged geographically remote from and in an operating environment of the autonomous agent; implementing a decisioning data buffer that includes the first stream of data from the onboard sensors and the second stream of data from the offboard sensors; generating current state data; generating/estimating intent data for each of one or more agents within the operating environment of the autonomous agent; identifying a plurality of candidate behavioral policies; and selecting and executing at least one of the plurality of candidate behavioral policies.
1. A system for operation of an autonomous agent, the system comprising:
a sensor system onboard the autonomous agent, wherein the sensor system is configured to generate a first data stream;
a remote infrastructure device configured to generate a second data stream;
a communication interface of the autonomous agent communicatively coupled to the remote infrastructure device;
a computing system distributed between at least the autonomous agent and the remote infrastructure device, wherein the computing system is configured to repeatedly:
calculate a probability value for each of a plurality of predetermined candidate behavioral policies using the first and second data streams;
select a behavioral policy from the plurality of predetermined candidate behavioral policies, based on a set of criteria comprising selecting a behavioral policy having the highest probability value from the plurality of predetermined candidate behavioral policies; and
autonomously control the autonomous agent according to the behavioral policy.
2. The system of claim 1 , wherein the computing system is configured to buffer at least one of the first and second data streams.
3. The system of claim 1 , wherein the computing system is further configured to:
track a set of agents in an environment of the autonomous agent based on the first and second data streams; and
estimate an intent of each agent of the set, wherein the behavioral policy is selected based on the estimated intent of at least one agent of the set.
4. The system of claim 3 , wherein each agent of the set is tracked within a predetermined region based on the second data stream.
5. The system of claim 1 , wherein the computing system further comprises a second remote infrastructure device.
6. A method comprising:
determining a first stream of observation data at an autonomous agent;
determining a second stream of observation data with a remote device arranged offboard the autonomous agent and communicatively coupled to the autonomous agent; and
repeatedly:
calculating a probability value for each of a plurality of predetermined candidate behavioral policies using the first and second streams of observation data;
selecting a behavioral policy from the plurality of predetermined candidate behavioral policies, based on a set of criteria comprising selecting a behavioral policy having the highest probability value from the plurality of predetermined candidate behavioral policies; and
autonomously controlling the autonomous agent according to the behavioral policy.
7. The method of claim 6 , wherein the behavioral policy is selected using a distributed computing system comprising a node offboard the autonomous agent.
8. The method of claim 7 , wherein the node is located at the remote device.
9. The method of claim 6 , wherein the behavioral policy is selected using a computing system, wherein the method further comprises: buffering at least one of the first and second streams of observation data at the computing system.
10. The method of claim 9 , wherein the computing system comprises a node onboard the autonomous agent.
11. The method of claim 6 , wherein determining the second stream of observation data comprises receiving the second stream of observation data at a computing system of the autonomous agent.
12. The method of claim 11 , wherein the second stream of observation data is received based on the autonomous agent being within a communication proximity of the remote device.
13. The method of claim 6 , wherein the first and second streams of observation data cooperatively define an overlapping field of sensing.
14. The method of claim 6 , wherein the second stream of observation data is associated with a predetermined region of sensing.
15. The method of claim 6 , wherein the second stream of observation data comprises abstractions of objects located within a predetermined region of sensing.
16. The method of claim 15 , wherein the abstractions of objects comprise a dynamic object data.
17. The method of claim 6 , wherein the first stream of observation data is determined using a sensor suite of the autonomous agent, wherein the remote device comprises: a sensor and a processor configured to generate the second stream of observation data based on measurements collected by the sensor.
18. The method of claim 6 , further comprising:
tracking a set of agents in an environment of the autonomous agent based on the first and second streams of data; and
determining an intent estimation for each agent of the set, wherein the behavioral policy is selected based on the intent estimation of at least one agent of the set.