Abstract
A kinematically controlled robot with a finite control range travels in a plane. There is an unknown dynamic scalar field in the plane; only the field value at the current location is measured. The robot should be autonomously driven to the unsteady and deforming isoline associated with a pre-specified field value and to track this isoline afterwards. A way to this objective is offered on the basis of a novel kinematic control paradigm dissimilar from the classic idea to turn the velocity perpendicular to the gradient. Specifically, a new reactive navigation strategy is presented and shown to solve the problem by a rigorous nonlocal convergence result in a generic field; this strategy does not employ estimation of gradient or other derivative-like quantities. These findings are applied to the problem of circumnavigation of several unknown moving targets based on range-only measurements. It is required to approach the targets and then to escort them so that they are enclosed inside the robot's path and a pre-specified small distance to the currently nearest target is maintained. Applicability and performance of the proposed circumnavigation approach is demonstrated via mathematically rigorous proofs and intensive computer simulations.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have