Architecture for multiple interacting robot intelligences

• US 7,328,196 B2
• Filed: 12/31/2003
• Issued: 02/05/2008
• Est. Priority Date: 12/31/2003
• Status: Expired due to Fees

• Current Assignee: Vanderbilt University
• Original Assignee: Vanderbilt University
• Inventors: Peters, Richard Alan II
• Primary Examiner(s): STARKS, WILBERT L
• Application Number:
  US10/749,326
• Publication Number:
  US 20050149227A1
• Time in Patent Office: 1,497 Days
• Field of Search: 706/14, 706/45, 700/247, 700/245, 700/259, 715/716, 715/854, 701/23, 382/155
• US Class Current: 706/14
• CPC Class Codes: B25J 9/161 Hardware, e.g. neural networks, fuzzy logic, interfaces, processor
• CPC Class Codes: B25J 9/163 learning, adaptive, model based, rule based expert control
• CPC Class Codes: G05B 2219/39254 Behaviour controller, robot have feelings, learns behaviour
• CPC Class Codes: G05D 1/0088 characterized by the autonomous decision making process, e.g. artificial intelligence, predefined behaviours using knowledge based models G06N5/00
• CPC Class Codes: G05D 1/0274 using mapping information stored in a memory device navigation using map-matching G01C21/30
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/008 based on physical entities controlled by simulated intelligence so as to replicate intelligent life forms, e.g. based on robots replicating pets or humans in their appearance or behaviour

Abstract

An architecture for robot intelligence enables a robot to learn new behaviors and create new behavior sequences autonomously and interact with a dynamically changing environment. Sensory information is mapped onto a Sensory Ego-Sphere (SES) that rapidly identifies important changes in the environment and functions much like short term memory. Behaviors are stored in a DBAM that creates an active map from the robot'"'"'s current state to a goal state and functions much like long term memory. A dream state converts recent activities stored in the SES and creates or modifies behaviors in the DBAM.

Claims

1. A method for providing short-term memory for an adaptive autonomous robot, the robot comprising at least one processor configured to execute at least one agent program, each of the at least one agent program configured to interact independently with other agent programs and at least one sensory processing unit associated with the at least one agent program, the method comprising:

creating a plurality of directions, each of the plurality of directions originating at the robot and having a set of nearest-neighbor directions defining a neighborhood associated with the each of the plurality of directions; and

associating an event detected by the at least one sensory processing unit with one of the plurality of directions.

2. The method of claim 1 further comprising fusing a first event associated with one of the plurality of directions with a second event associated with the same one of the plurality of directions.

3. The method of claim 1 further comprising selecting the direction having the greatest number of events associated with that direction as a focus for the robot.

4. The method of claim 1 wherein the number of the plurality of directions is fixed.

5. The method of claim 1 wherein the number of the plurality of directions is selected such that an angle between one of the plurality of directions with one of its nearest-neighbor directions is less than 10°

.

6. The method of claim 1 wherein the plurality of directions are spaced such that each direction corresponds to a vertex of a geodesic sphere centered on the robot.

7. The method of claim 1 further comprising re-associating an event previously associated with a previous direction with another one of the plurality of directions based on the movement of the robot.