Researchers led by Julian Martinez have developed a way to coordinate several drones to track and map air pollution as it spreads, according to a paper on arXiv. Their adaptive Density-Driven Control framework steers the drones using predictions of how the pollutant will move and target densities that are updated as new readings come in.

Reference samples of the plume are moved using concentration gradients and wind estimates measured on the spot, so the drones keep an up-to-date picture of the pollution field. The drones send their local measurements to a central coordinator, which keeps one shared set of samples and assigns targets and paths.

In simulations with mild and strong winds, the method mapped the pollution better than the baselines tested. When strong, changing winds pushed the plume beyond a fixed monitoring area, the baseline methods lost track while the new method followed it.

Source: arXiv Robotics (cs.RO)

Background

Julian Martinez

Robotics researcher

Robotics researcher.