DocumentCode :
3564139
Title :
An evolution approach to optimize the morphology of a six legged-wheeled hybrid mobile robot
Author :
Lim, S.H. ; Teo, J.
Author_Institution :
Evolutionary Comput. Lab., Univ. Malaysia Sabah, Kota Kinabalu, Malaysia
fYear :
2014
Firstpage :
1
Lastpage :
6
Abstract :
This paper proposed an evolution method in designing the morphology of a six legged-wheeled hybrid mobile robot which has the ability to climb over obstacles. A single objective evolution algorithm has been proposed to obtain an optimized morphology of a six legged-wheeled hybrid mobile robot with a smaller body while having best performance to climb over obstacles without collision. Results show that a fitter robot is obtained which is smaller and has a better performance in climbing obstacles than a manual hand-designed robot. The fittest robot obtained from the evolution has a very significant improvement in its fitness score compared to the manual hand-designed robot. In future, multi-objective evolution algorithm will be intended to be integrated into the evolution process to obtain a set of robot from the Pareto optimal solutions. Besides that, co-evolution of both of the controller and morphology will be investigated in order to have a more comprehensive evolution process.
Keywords :
Pareto optimisation; collision avoidance; control system synthesis; evolutionary computation; mobile robots; Pareto optimal solutions; collision avoidance; comprehensive evolution process; controller-morphology co-evolution; evolution process; fitness score; multiobjective evolution algorithm; obstacle climbing ability; single-objective evolution algorithm; six-legged-wheeled hybrid mobile robot morphology design; six-legged-wheeled hybrid mobile robot morphology optimization; Collision avoidance; Legged locomotion; Robot kinematics; Robot sensing systems; Wheels;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Computational Science and Technology (ICCST), 2014 International Conference on
Type :
conf
DOI :
10.1109/ICCST.2014.7045201
Filename :
7045201
Link To Document :
بازگشت