Adaptive Trajectory Generation Under Velocity Constraints Using Dynamical Movement Primitives

Abstract

This letter addresses the problem of adaptive trajectory generation under velocity constraints while preserving the desired path shape. We adopt the framework of Dynamical Movement Primitives (DMPs) for trajectory modeling that offers useful features for online trajectory modulation using spatial and temporal parameter adaptation. This letter proposes a temporal coupling for DMPs based on a repulsive potential that keeps the velocity within predefined limits while inherently ensuring path shape invariance. The performance and stability of the proposed method are proven by Lyapunov-like arguments, and simulations are presented to verify the result.