Biomimetic Navigation of Non-holonomic Underactuated Robots for Non-gradient Based Seeking and Tracking Extended Sources

A robot of the Dubins-car type travels in a planar workspace that hosts an unknown scalar field. Two sensors supply the robot with the point-wise field measurements. The paper presents a new biologically inspired control strategy that drives the robot to the set of the field maximizers, with particular attention to the case where they form a whole curve. This strategy is hybrid, operates in two discrete modes, and uses sliding mode control technique. If the maximizers form a whole curve, that strategy also endows the robot with the capacity of tracking this curve. The proposed strategy is justified by a mathematically rigorous non-local convergence result and via computer simulation tests. It does not use gradient estimation and is non-demanding with respect to computation and motion.