Recently, there has been an increased interest in the deployment of continuum robots in unstructured and challenging environments. However, the application of the state-of-the-art motion planning strategies, that have been developed for rigid robots, could be challenging in continuum robots. This, in fact, is due to the compliance that continuum robots possess besides their increased number of degrees of freedom. In this paper, a Demonstration Guided Pose Planning (DGPP) technique is proposed to learn and subsequently plan for spatial point-to-point motions for multi-section continuum robots. Motion demonstrations, including position and orientation, are collected from a human via a flexible input interface that is developed to command the continuum robot intuitively via teleoperation. A dynamic model based on Euler-Lagrange formalism is derived for a two-section continuum robot to be considered while planning for the robot motions. Meanwhile, a Proportional-Derivative (PD) computed torque controller with a Model Reference Adaptive Kinematic Control (MRAKC) scheme are developed to ensure the tracking performance against system uncertainties and disturbances. Also, the system stability analysis based on Lyapunov quadratic equation is proven. Simulation results prove that the proposed DGPP approach, along with the developed control scheme, have the ability to learn, generalize and reproduce spatial motions for a two-section continuum robot while avoiding both static and dynamic obstacles that could exist in the environments.
ASJC Scopus subject areas
- コンピュータ サイエンス（全般）