Robot Motion Planning Integrating Planning Strategies and Learning Methods

Robot motion planning in a dynamic cluttered workspace requires the capability of dealing with obstacles and deadlock situations. The paper analyzes situations where the robot is considered with its shape and size and it can only perceive the space through its local sensors. The robot explores the space using a planner based on an arti cial potential eld and incrementally learns a fast way to escape or prevent deadlock situations using a combination of sensor perceptions, eld information and planner experience. The knowledge acquired is a high-level network useful for avoiding deadlock areas consisting of local minimum nodes, backtracking nodes and subgoal nodes.

• Type: Conference
• Year: 1994
• Where: AIPS
• Authors: Luca Maria Gambardella, Cristina Versino
• Keywords: AIPS 1994, Artificial Intelligence, Deadlock Situations, Dynamic Cluttered Workspace, Paper Analyzes Situations