Hybrid reinforcement learning for autonomous driving
AssigneeQUALCOMM Incorporated
InventorsMohammad Naghshvar, Ahmed Kamel Sadek, Auke Joris Wiggers
Life of this patent
Estimating the term…
Abstract
A method includes determining a current state of an environment of an autonomous agent, such as a vehicle. The method also includes determining, via a first neural network, a set of actions based on the current state. The method further includes determining whether further analysis of the set of actions is desired. The method selects an action from the set of actions using a model-based solution based on a reward and a risk of the action when further analysis is desired. The method also includes selecting the action from the set of actions according to a metric when further analysis is not desired. The method controls the autonomous agent to perform the selected action.
Claims20
A processor-implemented method, comprising:
- determining a current state of an environment of an autonomous agent;
- determining, via a first neural network, a set of actions based on the current state;
- selecting an action from the set of actions using a model-based solution based on a reward for the action;
- and
- controlling the autonomous agent to perform the selected action.
An apparatus comprising:
- a memory;
- and
- at least one processor coupled to the memory, the at least one processor configured to:
- determine a current state of an environment of an autonomous agent;
- determine, via a first neural network, a set of actions based on the current state;
- select an action from the set of actions using a model-based solution based on a reward for the action;
- and
- control the autonomous agent to perform the selected action.
A non-transitory computer-readable medium having program code recorded thereon, the program code executed by a processor and comprising:
- program code to determine a current state of an environment of an autonomous agent;
- program code to determine, via a first neural network, a set of actions based on the current state;
- program code to select an action from the set of actions using a model-based solution based on a reward for the action;
- and
- program code to control the autonomous agent to perform the selected action.
Dependencies are inferred from the claim text.