Abstract
Inverse kinematics is one of the most important and complicated problems in robotics, and there is almost no exact analytical solution for this problem. Alternatively, with significant growth in machine learning techniques in recent decades, numerical methods are widely being used to solve this problem. This article aims to present a novel application of two
powerful meta-heuristic optimization algorithms including cuckoo optimization algorithm and imperialist competitive algorithm to solve robotic manipulators’ inverse kinematics problem for the first time. Recently, these two algorithms have been used to solve several problems in different majors more efficiently in comparison with other well-known algorithms.
To validate the efficiency of proposed approaches and suggest them as the preferred numerical methods to solve this problem, a comprehensive study has been done on performance of almost all recently used numerical methods to solve the same problem including genetic algorithm, particle swarm optimization, harmony search and differential evolution algorithms as well as adaptive neuro-fuzzy inference system and two artificial neural networks (multilayer perceptron and radial basis function). Simulations have been performed on an anthropomorphic arm with spherical wrist manipulator case study because of its complex model and degree of freedom. Proposed approach can be included in
computer-aided manufacturing packages to increase robotic manipulators’ efficiencies.