Trajectory Tracking Solution of a Robotic Arm Based on Optimized ANN

Abstract

The inverse kinematics solution of the serial robots has always been a challenge and needs complex calculations, while it's forward kinematics is easy. Therefore, a suitable and accurate manner for inverse kinematics problem (IKP) of the serial robots is very important due to different applications of trajectory planning of serial robots in many fields of robotics. In this paper, an optimized feed-forward artificial neural network (ANN) is applied to overcome the IKP of a planar arm with three-link, which is used to create desired trajectories in the two dimensional (2D) Cartesian space. The main target of this study is increasing the performance of the ANN using an optimization algorithm to find the best architecture of the network. Results show that the optimized network can manage the manipulator movements precisely to the desired position.