A novel framework for generalizing dynamic movement primitives under kinematic constraints

Abstract In this work, we propose a novel framework for generalizing desired trajectory pattern, encoded using Dynamic Movement Primitives (DMP), subject to kinematic constraints. DMP have been extensively used in robotics encoding and reproducing behaviours, thanks their generalization, stability robustness properties. However, incorporating constraints has not yet fully addressed. To end, design an optimization framework, based on the formulation from our previous patterns, with constraints, considering also time-varying target time duration, via-point obstacle Simulations highlight these properties comparisons are drawn other approaches enforcing DMP. The usefulness applicability of proposed is showcased experimental scenarios, including handover, where duration vary, placing obstacles dynamically introduced scene.