Boomerang algorithm based on swarm optimization for inverse kinematics of 6 DOF open chain manipulators

In this study, a feasible swarm intelligence algorithm is proposed that computes the inverse kinematics solution of 6 degree of freedom (DOF) industrial robot arms, which are frequently used in industrial and medical applications. The proposed algorithm is named as Boomerang algorithm due to its recursive structure. The proposed algorithm aims to reduce the computation time to feasible levels without increasing the position and orientation errors. In order to reduce the computational time in swarm optimization algorithms and increase feasibility, an alternative definition method was used instead of the DH method in defining the robot arm kinematic configuration. The effect of the proposed alternative definition method in reducing the computational time is presented through example inverse kinematic analysis. The proposed algorithm was compared with 3 different particle swarm optimization (PSO) variants that include orientation in the inverse kinematic solution of 6 DOF robot arms. Comparative simulation studies were carried out with 20 randomly selected position and orientation data from the workspaces of PUMA 560 and ABB IRB120 manipulators to measure performance of the algorithms. Using the error and computation time values obtained from the simulation results, the algorithms are compared using the Wilcoxon nonparametric statistical test. When the simulation results are analysed by considering the calculation time, positioning accuracy and solution finding rates, it is seen that the Boomerang algorithm is more feasible than the other PSO variants. Verification of the simulation results, and the physical applications were carried out with the ABB IRB120 6 DOF robot arm. Simulation studies and experimental studies showed that the proposed algorithm may be an efficient method for inverse kinematics of time-critical applications.