Event
Authors propose spatiotemporal GP-KF framework for multi-robot informative path planning
1 reports · 1 sources
Muqing Cao and other authors propose a grid-based spatiotemporal Gaussian process-Kalman filter framework for learned multi-robot informative path planning. The framework represents anonymous target presence with a single latent variable field on a discrete workspace grid, rather than maintaining a separate Gaussian process for each target. The authors claim that in simulation, the centralized deployment improves target visitation compared with learned methods and classical auction/coverage baselines, and reduces posterior target uncertainty by about 20%; a real two-drone experiment completed multi-robot search over an outdoor field of more than 7,000 square meters, showing that the method can transfer to real scenarios. Under decentralized deployment, each robot maintains its own mapper, and robots exchange fixed-size belief summaries rather than raw measurement data.
Report timeline
- Learning-Based Informative Path Planning for Multi-Robot Systems: A Spatio-Temporal Gaussian Process Kalman Filter Framework
The authors propose a grid-based spatio-temporal Gaussian process-Kalman filter framework for learning-based multi-robot informative path planning, representing anonymous target presence as a single latent variable field over a discrete workspace grid rather than maintaining one Gaussian process per target.
arXiv cs.RO ·