Navigation-Aware Path Planning and Multi-Agent Coordination in Challenging Environments
Kristen Michaelson, Manan Gandhi, Renato Zanetti · 2025
Effective information-gathering is crucial for teams of agents operating in challenging environments. Traditional path planning methods may fail to produce sufficiently informative trajectories. This work presents a cost function for optimal global planning that captures the total navigation uncertainty over the course of the trajectory. The cost metric, based on linear covariance analysis (LinCov), induces navigation-friendly behaviors by drawing agents into regions where they can collect informative measurements. This idea is extended for multi-agent planning: one agent acts a a moving beacon, providing positioning information to other agents operating within range.