Abstract: A 6-DOF Stautli industrial robot manipulator of TX60 type is regarded as the research object, and then nonlinear equations are constructed to solve its inverse positional problems by means of kinematics analysis. Taking the minimizing pose error of the end-effector as the objective function, this study transforms the nonlinear equations into an unconstrained optimization model and employs a differential evolution (DE) algorithm to solve this problem. To overcome the defects of the simple DE algorithm that is difficult to keep balance between the convergence accuracy and the calculation reliability, an adaptive division complex DE (ADCDE) algorithm is presented to enhance its optimizing performance. This proposed algorithm divides adaptively the population into exploitation and exploration subgroups, furthermore, the former and latter apply the DE/best/2/bin strategy and the complex operator to generate mutation individuals, respectively. With a study case for engineering application, the ADCDE algorithm is implemented to find the inverse positional solutions of a serial robot manipulator. Simulation results indicate that the presented algorithm outperforms the compared algorithms in term of computational precision and reliability, and the inverse solutions with high precision and all possible ones can be obtained by this algorithm with light calculation cost. Moreover, these results also verify the effectiveness and feasibility of this approach.

Key words: robot manipulator;inverse positional analysis;differential evolution algorithm;complex method;adaptive division