Multi-Robot Source Seeking and Field Reconstruction of Spatial-Temporal Varying Fields

To effectively respond to environmental disasters, real-time monitoring and pollution control rely heavily on the ability to estimate, predict, and reconstruct constantly changing environmental conditions across different locations. In this research, we develop a multi-robot source-seeking and field reconstruction framework that enables a multi-robot group to detect multiple sources within a spatial-temporal varying field and reconstruct the field in real time. The strategy consists of two parts: source seeking and field reconstruction. The source-seeking part features a gradient-based source-seeking controller that directs the multi-robot group toward local sources and a destination selection algorithm that navigates the multi-robot formation beyond the local maximal. The field reconstruction part reconstructs the spatial-temporal varying field in real time using the limited measurements taken by the robots. We validate the strategy in simulations.