Contactless Surface Following with Acceleration Limits: Enhancing Robot Manipulator Performance through Model Predictive Control

In robotics, automating surface following is desirable for applications such as spray painting, inspection, and surface finishing. However, current surface following approaches focus on in-contact applications, where the robot is restricted to operate within a slow dynamic range. We introduce an approach tailored for contactless surface following, which leverages Model Predictive Control (MPC) to better exploit the full dynamic range of the robot while taking its acceleration limits into account. Our formulation introduces a surface model based on Radial Basis Function Networks (RBFN), and we show that when it is combined with local sensing and a surface estimator, it achieves significantly better performance compared to state-of-the-art approaches which rely on a local quadratic model. The approach was validated through extensive simulations, and it was found to reliably compute feasible control inputs within realtime. Through this research we aim to enable less conservative surface-following behaviour and bring MPC closer to real-life industrial applications in robotics.