A new approach to improve the parameter estimation accuracy in robotic manipulators using a multi-objective output error identification technique

The research behind this article primarily concerns the development of mobile robots for nuclear decommissioning. The robotic platform under study has dual, seven-function, hydraulically actuated manipulators, for which the authors are developing a vision based, assisted teleoperation interface for common decommissioning tasks such as pipe cutting. However, to improve safety, task execution speed and operator training-time, high performance control of the nonlinear manipulator dynamics is required. Hence, the present article focuses on an associated dynamic model, and addresses the challenging generic task of parameter estimation for a highly non-convex and nonlinear system. A novel approach for estimation of the fundamental parameters of the manipulator, based on the idea of multi-objectivization, is proposed. Here, a single objective output error identification problem is converted into a multi-objective optimization problem. This is solved using a multi-objective genetic algorithm with non-dominated sorting. Numerical and experimental results using the nuclear decommissioning robot, show that the performance of the proposed approach, in terms of both the output error index and the accuracy of the estimated parameters, is superior to the previously studied single-objective identification problem.