Prediction Methods of Joint Control Torque Mechatronics Robots Based on GA Algorithms

Abstract

An optimal control strategy is implemented using a genetic algorithm (GA) to operate a machine vision-equipped tracker robot. First, the governing dynamic equations of the robot system are retrieved, and then the kinematic connections are collected in accordance with the requirements of the camera. The obtained open-loop robot model is then validated through simulation and experimentation. Initial robotic system control is implemented using a computed torque control algorithm with a standard tuning strategy. The harsh non-linearity of the robot controller, such as actuator saturation, makes it difficult to tune the aforementioned controller using standard tuning methods. As a result, this study proposes an improved optimal control method based on a genetic algorithm. Consequently, GA is utilised to acquire optimal values of control constants in accordance with an appropriate cost function. As such, we undertake a simulation study to demonstrate the benefits of the GA-based controller over the conventional one. Additionally, both the mathematical and experimental foundations for the linking strategy used to connect the vision subsystem to the robot controller are provided. The performance of the proposed visual tracker robot is then studied as it follows an object along a circular and a butterfly-shaped route in camera information. The experimental findings verify the robustness of the control system and demonstrate that the visual robot equipped with the optimal controller is successful in following the target trajectories.