A two-stage distributed receding horizon control for persistent monitoring tasks with monitoring count requirements

This paper investigates the problem of multi-agent persistent monitoring with monitoring count requirements. A team of agents is tasked with monitoring multiple targets to minimise mean overall target state uncertainties, while simultaneously satisfying the monitoring count requirements of specific targets within predetermined time intervals. The inherent demands of assigning different monitoring tasks to each agent and determining the appropriate execution times introduce significant complexities, making it difficult to meet the monitoring count requirements. To address this, a persistence predicate based on Signal Temporal Logic (STL) specifications is introduced, which allows monitoring count requirements to be concisely expressed as multiple sub-STL requirements. A two-stage distributed receding horizon control (DRHC) algorithm is then proposed to optimise the agent trajectory, including their arrival and dwell times on the targets. In the first stage, each sub-STL task is dynamically reallocated within the DRHC framework to ensure feasibility. In the second stage, a finite horizon DRHC problem is formulated to optimise the monitoring performance while meeting the local monitoring count requirements. Agents solve their problems sequentially to obtain controllers that optimise overall monitoring performance while ensuring the satisfaction of global STL tasks. Simulation results are provided to demonstrate the effectiveness of the proposed approach.