Intelligent control for accurate fast response and minimum energy of motion for industrial robotic manipulator

摘要

This work proposes an approach to optimise the performance of a six degrees-of-freedom (6-DOF) robotic manipulator. The focus is on achieving a balance between three key objectives: rapid response speed, minimal positioning error, and reduced energy consumption during movement. The methodology employs a two-phase approach. First, a kinematic model is established using the Denavit-Hartenberg convention. Subsequently, a grey wolf optimiser (GWO) identifies optimal joint configurations for diverse target locations within the workspace. These optimal configurations serve as training data for a forward neural network (FNN) model, enabling it to predict optimal joint angles for future tasks. The proposed method demonstrates exceptional capability in precisely positioning the manipulator at desired locations within a short timeframe (0.01 sec average) while maintaining high accuracy (0.0056 mean square error (MSE) average) and achieving significant energy savings (70% average reduction). This approach presents a promising solution for enhancing the overall performance of 6-DOF robotic manipulators.