VoltaResBot: A Machine Learning Model for Optimal Energy Management in Multi-Component Robotic Systems Integrated with Photovoltaics, and Storages

Abstract

Predictive energy management models considerably advance financial and environmental analyses of industrial robotic manipulator energy consumption. As industries increasingly prioritize sustainability and cost-effectiveness, optimizing energy utilization becomes crucial. Consequently, this paper presents VoltaResBot, a machine learning-driven predictive techno-economic model for optimal energy management in multi-component robotic systems integrated with photovoltaics (PVs), electricity grid, and storage (ESS). VoltaResBot utilizes Support Vector Machines (SVM) to predict and optimize energy consumption, facilitating precise control of a 6 Degrees of Freedom (6 DoF) robotic manipulator. The findings of VoltaResBot indicate its effectiveness in achieving optimal energy utilization, significantly reducing environmental. Key Performance Indicators (KPIs) employed in the evaluation include Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Mean Squared Error (MSE), and R-squared ($\mathbf{R}^{\mathbf{2}}$), presenting the model’s accuracy. Comparative analyses with traditional machine learning models further demonstrate the superior performance of VoltaResBot.