Chaotic artificial bee colony approach to step planning of maintaining balance for quadruped robot. Issue 2 (3rd June 2014)

• Record Type: Journal Article
• Authors: Luo, Qinan; Duan, Haibin
• Is Part Of: International journal of intelligent computing and cybernetics. Volume 7:Issue 2(2014)
• Page Start: 175
• Page End: 191
• Publication Date: 2014-06-03
• Subjects: Optimal control -- Robotics -- Chaotic -- Step planning -- Artificial bee colony (ABC) -- Quadruped robot; Artificial intelligence -- Periodicals; Cybernetics -- Periodicals; 006.3
• DOI: 10.1108/IJICC-05-2013-0028
• Languages: English
• ISSNs: 1756-378X

Abstract

Abstract : Purpose: – Artificial bee colony (ABC) algorithm is a relatively new optimization method inspired by the herd behavior of honey bees, which shows quite intelligence. The purpose of this paper is to propose an improved ABC optimization algorithm based on chaos theory for solving the push recovery problem of a quadruped robot, which can tune the controller parameters based on its search mechanism. ADAMS simulation environment is adopted to implement the proposed scheme for the quadruped robot. Design/methodology/approach: – Maintaining balance is a rather complicated global optimum problem for a quadruped robot which is about seeking a foot contact point prevents itself from falling down. To ensure the stability of the intelligent robot control system, the intelligent optimization method is employed. The proposed chaotic artificial bee colony (CABC) algorithm is based on basic ABC, and a chaotic mechanism is used to help the algorithm to jump out of the local optimum as well as finding the optimal parameters. The implementation procedure of our proposed chaotic ABC approach is described in detail. Findings: – The proposed CABC method is applied to a quadruped robot in ADAMS simulator. Using the CABC to implement, the quadruped robot can work smoothly under the interference. A comparison among the basic ABC and CABC is made. Experimental results verify a better trajectory tracking response can be achieved by the proposed CABC method after control parameters training. Practical implications: – The proposed CABC algorithm can be easily applied to practice and can steer the robot during walking, which will considerably increase the autonomy of the robot. Originality/value: – The proposed CABC approach is interesting for the optimization of a control scheme for quadruped robot. A parameter training methodology, using the presented intelligent algorithm is proposed to increase the learning capability. The experimental results verify the system stabilization, favorable performance and no chattering phenomena can be achieved by using the proposed CABC algorithm. And, the proposed CABC methodology can be easily extended to other applications.