Path Following with Stable and Unstable Modes Subject to Time-Varying Dwell-Time Conditions

Abstract Systems are often tasked with operating in modes where state feedback is available intermittently. The analysis of such systems involves analyzing the behavior of individual subsystems and the time that each subsystem is active, i.e., dwell-time conditions. Often, these dwell-time conditions are conservative, potentially limiting the performance of the overall system. In an effort to reduce conservativeness of the dwell-time condition, an adaptive data-driven extremum seeking (ESC) method is used to develop a time-varying dwell-time condition. Specifically, the ESC drives the evolution of the dwell-time condition to a less conservative dwell-time condition while simultaneously ensuring stability of the overall system. Simulations demonstrate a nearly threefold increase in a maximum dwell-time that results in significant changes to the behavior of an agent tasked with following a path outside a feedback region.