DocumentCode :
2070406
Title :
Using software agents to avoid collisions among multiple robots
Author :
Jäger, Markus
Author_Institution :
Corporate Technology, Inf. & Commun., Siemens AG, Munich, Germany
fYear :
2001
fDate :
7-9 Nov 2001
Firstpage :
50
Lastpage :
57
Abstract :
The paper describes a method where collaborative software agents are used to coordinate the independently planned trajectories of multiple mobile robots to avoid collisions and deadlocks among them. Whenever the distance between two robots drops below a certain value, the agents exchange information about the planned trajectories of the robots and determine whether they are in danger of a collision. If a possible collision is detected, the agents monitor the robot movements and, if necessary, insert idle times between certain segments of the trajectories in order to avoid the collision. Deadlocks among two or more robots occur if a number of robots block each other in such a way that none of them is able to continue along its trajectory without causing a collision. These deadlocks are reliably detected by the agents. After a deadlock is detected, alternative trajectories for each of the involved robots are successively planned until the deadlock is resolved. The agents use a combination of three fully distributed algorithms to reliably solve the task. They do not use any global synchronization
Keywords :
collision avoidance; distributed algorithms; intelligent control; mobile robots; software agents; alternative trajectories; collaborative software agents; collision avoidance; deadlocks; fully distributed algorithms; idle times; independently planned trajectories; information exchange; mobile robots; multiple robots; planned trajectories; robot movement monitoring; Collaboration; Collision avoidance; Distributed algorithms; Mobile communication; Mobile robots; Monitoring; Robot kinematics; Software agents; System recovery; Trajectory;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Tools with Artificial Intelligence, Proceedings of the 13th International Conference on
Conference_Location :
Dallas, TX
Print_ISBN :
0-7695-1417-0
Type :
conf
DOI :
10.1109/ICTAI.2001.974448
Filename :
974448
Link To Document :
بازگشت