Multi-3 adaptive coordination of robots

Background: Multi-robot systems (MRSs) are gaining interest as an effective means for addressing complex tasks in various domains. Objective: This paper focuses on three multiplicity problems in MRSs to improve flexibility and resource utilization and reduce idle times or busy-waiting when possible: (i) enabling multi-purpose robots to execute multiple missions at a time, (ii) modeling MRS missions with multi-instance tasks, i.e., tasks repeated for multiple locations or objects, and (iii) enabling dynamic task assignment to improve resource utilization. Methods: This work builds on the limitations of the literature to enable the execution of concurrent missions through a fleet of robots and, within each mission, multiple instances of the same task. The solution is experimentally validated to check its effectiveness in coordinating robots and reducing mission completion time, idle time, and battery consumption. Results: We present Multi-3, a framework for executing multiple missions (multi-missions) with multiple instances of the same task (multi-instance) using multiple robots (multi-robots). At the specification level, it enables the definition of missions with multi-instance tasks. At the execution level, the runtime support permits the execution of multiple missions at a time, while the adaptive coordination of the involved robots allows them to dynamically switch across tasks of multiple missions or multiple task instances. Conclusion: We validate the solution at work on a set of multi-robot systems scenarios in different domains. The results show that Multi-3 is viable, reduces the robots’ idle time, and effectively allows the execution of multiple missions at a time.