Method and system for data-driven and modular decision making and trajectory generation of an autonomous agent
A system for data-driven, modular decision making and trajectory generation includes a computing system. A method for data-driven, modular decision making and trajectory generation includes: receiving a set of inputs; selecting a learning module such as a deep decision network and/or a deep trajectory network from a set of learning modules; producing an output based on the learning module; repeating any or all of the above processes; and/or any other suitable processes. Additionally or alternatively, the method can include training any or all of the learning modules; validating one or more outputs; and/or any other suitable processes and/or combination of processes.
1 . A method for operating a vehicle, the method comprising:
with a set of decision-making modules:
with a first of the set of decision-making modules, processing a set of sensor inputs to produce an intermediate output comprising an action;
based on the intermediate output, and with a second of the set of decision-making modules, producing a trajectory; and
operating the vehicle according to the trajectory.
2 . The method of claim 1 , wherein the set of decision-making modules comprises multiple learned models.
3 . The method of claim 2 , wherein the first of the set of decision-making modules is part of a first subset of decision-making modules, the first subset comprising multiple decision-making modules.
4 . The method of claim 2 , wherein the first and second decision-making modules are separate and distinct.
5 . The method of claim 4 , wherein the first and second decision-making modules are configured to receive different input types and produce different output types.
6 . The method of claim 4 , further comprising selecting the second decision-making module based on and in response to producing the intermediate output.
7 . The method of claim 1 , wherein the vehicle is an autonomous truck.
8 . The method of claim 1 , wherein producing the trajectory comprises selecting the second decision-making module based on the action.
9 . The method of claim 1 , wherein processing the set of sensor inputs is performed at a first time during a vehicle trip, wherein the method further comprises:
with a third of the set of decision-making modules, at a second time during the vehicle trip later than the first time, processing a second set of sensor inputs to produce a second intermediate output;
based on the second intermediate output, and with a fourth of the set of decision-making modules, producing a second trajectory; and
controlling the vehicle according to the second trajectory.
10 . The method of claim 9 , wherein:
the first and third decision-making modules are separate and distinct relative to each other; and
the second and fourth decision-making modules are separate and distinct relative to each other.
11 . A system for operating a vehicle, the system comprising:
a set of decision-making modules;
a processing subsystem comprising a set of processors configured to:
with a first of the set of decision-making modules, process a set of sensor inputs to produce an intermediate output comprising an action;
based on the intermediate output, and with a second of the set of decision-making modules, produce a trajectory; and
a control subsystem comprising a set of vehicle controllers configured to operate the vehicle according to the trajectory.
12 . The system of claim 11 , wherein the set of decision-making modules comprises multiple learned models.
13 . The system of claim 12 , wherein the first of the set of decision-making modules is part of a first subset of decision-making modules, the first subset comprising multiple decision-making modules.
14 . The system of claim 12 , wherein the first and second decision-making modules are separate and distinct.
15 . The system of claim 14 , wherein the first and second decision-making modules are configured to receive different input types and produce different output types.
16 . The system of claim 14 , wherein the second decision-making module is selected based on and in response to producing the intermediate output.
17 . The system of claim 11 , wherein the vehicle is an autonomous truck.
18 . The system of claim 11 , wherein processing the set of sensor inputs is performed at a first time during a vehicle trip, wherein:
the processing subsystem is further configured to:
with a third of the set of decision-making modules, at a second time during the vehicle trip later than the first time, process a second set of sensor inputs to produce a second intermediate output;
based on the second intermediate output, and with a fourth of the set of decision-making modules, produce a second trajectory; and
the control subsystem is further configured to operate the vehicle according to the second trajectory.
19 . The system of claim 18 , wherein:
the first and third decision-making modules are separate and distinct relative to each other; and
the second and fourth decision-making modules are separate and distinct relative to each other.