Global Temporal Logic Control Synthesis for Multiagent Systems With Time and Space Margin

This article considers the multiagent control problem under global temporal logic tasks, where agents possess heterogeneous capabilities. The global temporal logic task imposes requirements such as frequency and agent capabilities on the completion of signal temporal logic (STL) tasks, while also encompassing coupling relationships between agents. The complex formulation makes it challenging to achieve efficient task assignments with desired spatial-temporal control performance. To tackle this issue, a global language called extended capability temporal logic plus (ECaTL+) is first introduced for task formulation. The robustness measure for ECaTL+ is formally provided from both space and time perspectives. A hierarchical model predictive control (MPC) framework is then proposed to achieve task-satisfied motion planning with time and space margins, along with efficient computation. In the upper layer, a centralized MPC is formulated to allocate ECaTL+ tasks for time robustness maximization. Two mixed-integer linear programming (MILP) encoding methods are provided: one is complete, while the other employs formula simplification to save computation costs. Building upon the task assignment results, the lower layer focuses on realizing path planning for each agent by maximizing space robustness within the framework of iterative distributed MPC. Simulations and experiments are conducted to demonstrate the efficiency of the proposed algorithm.