TY - CHAP
T1 - Understanding the behaviour of reactive robots in a patrol task by analysing their trajectories
AU - Kumar, Swagat
AU - Thach, Nguyen Huy
AU - Suzuki, Einoshin
PY - 2010/11/1
Y1 - 2010/11/1
N2 - n this paper, we present our analysis of trajectories for understanding the behaviour of reactive robots in a patrol task. Reactive robots are considered as one of the main applications of agent technologies and are hence one of the major resources of new challenges for the technologies. The task of our 4 real robots, each of which is equipped with 3 infrared sensors and an identical controller, is to patrol in a closed workspace composed of 64 cells. The performance of the robots is measured through a patrol rate, which is essentially the average rate of the cells visited by at least one robot in a given time interval. We test 15 kinds of controllers and try to understand their behaviours by analysing their trajectories filmed by an overhead camera. The logs are analysed using three different features which are relevant to the collective task and the inferences obtained from this analysis are related to the properties of the controllers. We believe that the insights obtained through this analysis would be useful for practitioners and researchers of reactive, distributed robots and related agent technologies.
AB - n this paper, we present our analysis of trajectories for understanding the behaviour of reactive robots in a patrol task. Reactive robots are considered as one of the main applications of agent technologies and are hence one of the major resources of new challenges for the technologies. The task of our 4 real robots, each of which is equipped with 3 infrared sensors and an identical controller, is to patrol in a closed workspace composed of 64 cells. The performance of the robots is measured through a patrol rate, which is essentially the average rate of the cells visited by at least one robot in a given time interval. We test 15 kinds of controllers and try to understand their behaviours by analysing their trajectories filmed by an overhead camera. The logs are analysed using three different features which are relevant to the collective task and the inferences obtained from this analysis are related to the properties of the controllers. We believe that the insights obtained through this analysis would be useful for practitioners and researchers of reactive, distributed robots and related agent technologies.
UR - http://www.mendeley.com/research/understanding-behaviour-reactive-robots-patrol-task-analysing-trajectories
U2 - 10.1109/WI-IAT.2010.250
DO - 10.1109/WI-IAT.2010.250
M3 - Chapter
SN - 9780769541914
T3 - Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010
SP - 56
EP - 63
BT - Proceedings - 2010 IEEE/WIC/ACM International Conference on Intelligent Agent Technology, IAT 2010
T2 - 2010 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology
Y2 - 31 August 2010 through 3 September 2010
ER -