Evolving Runge-Kutta-Gill RBF networks to estimate the dynamics of a multi-link manipulator

Proposes a method for identification of dynamics of a multi-link robot arm using Runge-Kutta-Gill neural networks (RKGNN). Shape adaptive radial basis function (RBF) neural networks have been employed with an evolutionary algorithm to optimize the shape parameters and the weights of the RKGNN. Due to the fact that the RKGNN can accurately grasp the changing rates of the states, this method can effectively be used for long term prediction of the states of the robot arm dynamics. Unlike in conventional methods, the proposed method can even be used without input torque information because a torque network is part of the functional network. This method can be proposed as an effective option for dynamics identification for manipulators with high degrees of freedom, as opposed to the derivation of dynamic equations and making additional hardware changes in the case of statistical parameter identification such as linear least-squares method