Development of a Demonstration-Guided Motion Planning for Multi-section Continuum Robots

Abstract

Recently, a considerable trend is growing towards engaging continuum robots in the navigation of confined environments. Due to their increased Degrees of Freedom (DOF), having a fully autonomous motion planning could be challenging in continuum robots. In this paper, a Demonstration-Guided Motion Planning (DGMP) framework is developed to teach continuum robots how to achieve point-point spatial motions from given demonstrations. A flexible rod is used as an input interface to demonstrate motions for the robot via teleoperation. The Dynamic Movement Primitives (DMP) framework is adopted to learn, reproduce and generalize the given demonstrations while avoiding novel obstacles that could exist in the environment. Meanwhile, a Model Reference Adaptive Controller (MRAC) is developed to ensure the robustness towards executing the motions generated from the DGMP by the robot. The developed approach is assessed over a simulated kinematic model of a two-section continuum robot. The results show evidence that the proposed DGMP is effective in generating and tracking spatial motions for continuum robots, which encourages further investigation towards planning complex motions in future.