Sequential monitoring for robot teams via temporal logic planning and persistent coverage control

Abstract

This paper proposes a multi-layer control framework for monitoring environments composed of subregions under Linear Temporal Logic (LTL) specifications. Task specifications described in natural language are translated into LTL formulas for temporal logic planning. An intermediate planning layer constructs a product automaton and determines a subregion monitoring sequence that satisfies the LTL constraints through shortest-path search. The transition cost incorporates both inter-region transition time and monitoring completion time estimated from the properties of time-varying Finite Time Convergence Control Barrier Functions (FCBFs). A lower control layer employs FCBFs to perform both inter-region transitions and region monitoring. Simulation results demonstrate that the proposed framework successfully integrates high-level logical task specifications with low-level continuous control and generates robot team behaviors that satisfy temporal monitoring constraints specified in natural language.