Multi-Objective Optimization for Inverse Kinematics of Robotic Arm with Joint Configuration

The hardware configuration of modern robots is rapidly developing, yet the fundamental problem of inverse kinematics is still mostly based on analytical or numerical solutions applied to mainstream robotic arms. These traditional methods for inverse kinematics result in an infinite number of solutions when encountering singularities, leading to uncertainty in which solution the robotic arm should use. In this paper, the inverse kinematics of two collaborative robot arm models are investigated using a multi-objective optimization weighted sum method and a particle swarm optimization algorithm. The objective functions include the position and orientation errors of the end-effector and the total rotation amount of the robotic joints. Experimental results show that including a total rotation amount of robotic joints in the objective functions effectively reduces energy consumption. Finally, through simulations and physical model validations according to a series of trajectories obtained, the developed algorithm is proven to achieve the inverse kinematics solution within the allowable error range for engineering requirements.