Motion Planning for Point-to-Point Navigation of Spherical Robot Using Position Feedback

Abstract

Nonlinear dynamics, slip, and irregularities in the surface or the spherical hull are the major factors for errors in the spherical robot system model. We attempt a controller design based on the experimental observations on spherical robot motion and identify the primitive motions that are reliable. The objective of the controller is to use these primitive motions and position feedback to guarantee convergence to the destination position. It is shown that the robot converges to a bounded region at the destination, when the modeling and measurement errors are bounded. The control law is applied to the navigation of a pendulum type spherical robot. The proposed navigation technique is not specific to the nonlinear spherical robot model and gives advantage by using limited control commands. Experiments in indoor and outdoor settings support the theory presented in this work. Experiments also show that the proposed controller rejects disturbances due to imperfect fabrication and contact surfaces with changing frictional properties. In particular, the robot navigates to the vicinity of the destination within a known range.