Hybrid pso-newton-raphson algorithm for inverse kinematics problem in robotics

Abstract

Newton-Raphson method is a deterministic numerical method for solving a system of nonlinear equations. In robotics, it is used to solve inverse kinematics problems. In order to converge towards the optimal solution, the Newton-Raphson method requires a good initial value guess, which can be challenging to obtain. The Particle Swarm Optimization (PSO) algorithm is a stochastic optimization technique for solving nonlinear problems. The advantage of the PSO, in this case, is its ability to search a large amount of data. The PSO can narrow down potential solutions close to the optimal solution and use them as an initial guess for the Newton-Raphson method. Then, the Newton-Raphson method takes over and converges towards the desired optimal solution. In this paper, the feasibility of the hybrid PSO-Newton-Raphson algorithm for solution of robot inverse kinematics problem is investigated for a six-degree of freedom robot arm. All six joints of the robot arm are revolute. The cost function for the PSO algorithm is formed as a function of error between the desired and actual position of the robot arm end-effector. The numerical simulation is carried out to verify the applicability of the proposed concept.