Reconfiguration in Motion Planning of Single- and Multi-agent Systems under Infeasible Local LTL Specifications

Abstract—A reconfiguration method for the model-checking-based motion planning of single- and multi-agent systems under infeasible local LTL specifications is proposed. The method describes how to synthesize the motion plan that fulfills the infeasible task specification the most, and how the infeasible task specification is relaxed. The novelty is the introduction of a metric within the atomic proposition domain, and the relative weighting between the implementation cost of a motion plan and its distance to the original specification. For multi-agent systems, a dependency relation and relative priorities are incorporated when the tasks are assigned independently to each agent. Simulations are presented to illustrate the method. I