Predictive Multi-Robot Task Allocation for Radiation Monitoring
UAV and UGV teams dispatched across a thermosolar plant to map direct normal irradiance, with a receding-horizon allocation that anticipates where measurements will matter next.
In collaboration with University of Seville
A thermosolar plant’s output depends on direct normal irradiance, which drifts across the field as clouds move. Measuring it well means putting sensors in the right place before the interesting thing happens, not after.
Working with the University of Seville under Prof. J. M. Maestre, this project implemented a high-level allocation method for a heterogeneous team of UAVs and UGVs. The allocation runs over a receding horizon and predicts how the irradiance field will evolve, dispatching robots to the cells where a measurement will most reduce uncertainty. It was validated in ROS and Gazebo against the La Africana plant layout.
The work was published in Solar Energy (2023) and at the European Control Conference 2022.