Optimal Energy Consumption Planning for Industrial Robots using Stochastic Gradient Descent based Federated Learning

Due to the broad distribution as well as minimum energy efficiency, an energy-saving of the industrial robots’ attractions effectively as well as various approaches have arisen to optimize and energy consumption of the robots. Nevertheless, various dynamic as well as electrical parameters are unattainable because of the constraints of commercial industrial robots. Hence, this research proposes the energy consumption for optimal motion planning for industrial robot using Stochastic Gradient Descent based Federated Learning (SGD-FL) approach. An objective function for the energy minimization issues is presented according to the energy consumption and execution time of the robot arm configurations in its workplace for pick-up tasks. The experimental results of the proposed SGD-FL approach attain the energy consumption of 55.98% and execution time of 5.385s respectively. These results show that the proposed method obtains superior results when compared to the conventional methods like Genetic Algorithm with Particle Swarm Optimization Algorithm (GAPSO) and Simulated Annealing (SA).