Efficient reinforcement learning of navigation strategies in an autonomous robot
This paper describes a reinforcement connectionist learning architecture that allows an autonomous mobile robot to adapt to an unknown indoor environment. As a result, the robot learns efficient reactive navigation strategies. The robot learns on top of built-in reflexes. Also, the neural controller is a modular network which is built automatically. In this way, the robot is operational from the very start and improves its performance rapidly and incrementally as it safely explores the environment. These features make this learning robot's architecture very well suited to real-world applications. The paper also reports experimental results obtained with a real mobile robot that demonstrate the feasibility of this approach.
Bibliographic Reference: Article: Intelligent Robots and Systems (1994)
Record Number: 199511169 / Last updated on: 1995-08-23
Original language: en
Available languages: en