System and method for implementing reward based strategies for promoting exploration
AssigneeHonda Motor Co., Ltd.
InventorDavid Francis Isele
Life of this patent
Estimating the term…
Abstract
A system and method for implementing reward based strategies for promoting exploration that include receiving data associated with an agent environment of an ego agent and a target agent and receiving data associated with a dynamic operation of the ego agent and the target agent within the agent environment. The system and method also include implementing a reward function that is associated with exploration of at least one agent state within the agent environment. The system and method further include training a neural network with a novel unexplored agent state.
Claims20
A computer-implemented method for implementing reward based strategies for promoting exploration comprising:
- receiving data associated with an agent environment of an ego agent and a target agent;
- receiving data associated with a dynamic operation of the ego agent and the target agent within the agent environment;
- implementing a reward function that is associated with exploration of at least one agent state within the agent environment, wherein the reward function includes assigning at least one reward based on if the at least one agent state is a novel unexplored agent state or a previously explored agent state;
- and
- training a neural network with the novel unexplored agent state, wherein at least one simulation is processed to determine at least one additional novel unexplored agent state based on an analysis of at least one reward of the reward function.
A system for implementing reward based strategies for promoting exploration comprising:
- a memory storing instructions when executed by a processor cause the processor to:
- receive data associated with an agent environment of an ego agent and a target agent;
- receive data associated with a dynamic operation of the ego agent and the target agent within the agent environment;
- implement a reward function that is associated with exploration of at least one agent state within the agent environment, wherein the reward function includes assigning at least one reward based on if the at least one agent state is a novel unexplored agent state or a previously explored agent state;
- and
- train a neural network with the novel unexplored agent state, wherein at least one simulation is processed to determine at least one additional novel unexplored agent state based on an analysis of at least one reward of the reward function.
A non-transitory computer readable storage medium storing instructions that when executed by a computer, which includes a processor perform a method, the method comprising:
- receiving data associated with an agent environment of an ego agent and a target agent;
- receiving data associated with a dynamic operation of the ego agent and the target agent within the agent environment;
- implementing a reward function that is associated with exploration of at least one agent state within the agent environment, wherein the reward function includes assigning at least one reward based on if the at least one agent state is a novel unexplored agent state or a previously explored agent state;
- and
- training a neural network with the novel unexplored agent state, wherein at least one simulation is processed to determine at least one additional novel unexplored agent state based on an analysis of at least one reward of the reward function.
Dependencies are inferred from the claim text.