Improving Inverse Kinematic solutions of Robotic Systems for Accuracy and Speed Using Neural Network Ensembles

One of the significant problems in design and control a robot is robot’s kinematics, in which inverse kinematics is critical for high DoF robots. Although, many algorithms have been employed to solve this problem, the deficiency of efficient methods is observable. As the Complexity of robot increases, solving the inverse kinematics is difficult and computationally expensive. Traditional methods such as geometric, iterative and algebraic are inadequate, if the robot is more complex. As an alternative approach, neural networks have been widely used for inverse kinematic modeling. This paper investigates two neural network ensembles, i.e., averaging and mixture of experts, to solve the inverse kinematics for the robotic manipulators, then it is applied to a manipulator with 3 degrees of freedom as a case study. The obtained results from the methods are presented and analyzed in order to prove the efficiency of the proposed approach in comparison with analytical solutions.