Flexible Joint Manipulator

Flexible joint manipulators, robots with compliant joints, present a control challenge due to their inherent elasticity, impacting trajectory accuracy, especially at high speeds. Current research focuses on developing advanced control strategies, including model predictive control (MPC) leveraging singular perturbation theory and deep reinforcement learning (DRL) methods with reference correction or learned two-stage models incorporating future state prediction. These approaches aim to improve trajectory tracking accuracy and robustness against external disturbances, enabling safer and more precise human-robot collaboration in shared workspaces.