Multiple Demonstrations Based Trajectory Generation for Mobile Agent Using DMPs with T-S Fuzzy Model

Abstract

Dynamic Movement Primitives (DMPs), which represent movement plans based on a set of nonlinear differential equations with well-defined attractor dynamics, have been adapted for movement planning. In this paper, we present a trajectory generation method for mobile agent from multiple demonstration using DMPs. The original DMPs can be only used to learn from single demonstration. In order to adapt multiple demonstrations of different tasks, we use DMPs with Takagi-Sugeno fuzzy model to realize autonomous trajectory planning. The proposed scheme is capable to regenerate trajectory in the presence of obstacles even when the goal position is altered. A simulation example is given to confirm the effectiveness of the proposed scheme.