engleski

Simulation of intelligent robot behavior based on reinforcement learning and neural network approach

The paper presents a planning system which integrates the reinforcement learning method and a neural network approach with the aim to ensure autonomous robot behavior in unpredictable working conditions. The assumption is that the robot is a tabula rasa and has no knowledge of the work space structure. Initially, it has just basic strategic knowledge of searching for solutions, based on random attempts, and a built-in learning system. It explores the work space by a simple sensor system, learning on-line, and being rewarded for successful action or punished for the action which does not lead to the goal state. The reinforcement learning method is used here to evaluate robot behavior and to induce new or to improve the existing knowledge. The acquired action (task) plan is stored as experience which can be used in solving future similar problems. The recognition of similar problems is established on recognition of work space structure as a structural assignment problem.

automatic assembly, intelligent planning, robotics, learning

nije evidentirano