Advancements in Multi-Robot Informative Path Planning for Target Monitoring
A new framework enhances multi-robot path planning under uncertainty.
Recent research introduces a grid-based spatio-temporal GP-Kalman filtering framework aimed at improving multi-robot informative path planning (IPP) for persistent target monitoring. This method addresses challenges related to spatial uncertainty and communication constraints by representing target presence as a latent field over a discrete grid.
The framework allows for decentralized deployment, where each robot maintains its own mapper and shares compact belief summaries. This approach leads to a significant reduction in target uncertainty and enhances target visitation rates, as demonstrated in both simulation benchmarks and real-world UAV experiments.