A beta salp swarm algorithm meta-heuristic for inverse kinematics and optimization

This paper first reviews heuristic-based and bio-inspired contributions in inverse kinematics. A new inverse kinematics solver is then proposed based on beta distributed Salp Swarm Algorithm called β-SSA. The proposed algorithm is an alternative of the SSA algorithm where leading salps are distributed based on the beta function, enabling a better control of their repartition on the search space. The β -SSA inverse kinematics solver is named IK- β -SSA and can be considered as a generic framework. It uses a generic formulation of a forward kinematic model of a robotic system to retrieve its inverse solution. Inverse solution consists in obtaining a possible and feasible joint motions allow the robotic system to achieve a specific position while satisfying intrinsic constraints such as joints positions/ velocities limitations or path limitations. The β -SSA algorithm is first tested on a set of test functions and compared to nominal SSA prior to be applied to solve the inverse kinematics problem of the industrial robotic arm, Kuka Kr05-arc. The proposed method shows very competitive results when compared to classical SSA, QPSO, Bi-PSO, K-ABC and FA. The experimental results based on simulations and a Wilcoxon non-parametric statistical tests evidently show that the IK-β -SSA performs better than classical SSA, QPSO, Bi-PSO, K-ABC and FA for a single point inverse kinematics solution using a generic 8 Dof arm and the Kr05 industrial robot. For the path planning, a circular path tracking was investigated using the Kr05 robot and confirmed also that the β -SSA performs better than classical SSA, QPSO, Bi-PSO, K-ABC and FA.