A Local PSO-Based Algorithm for Cooperative Multi-UAV Pollution Source Localization

Recently, air pollution has grown significantly, and it can frequently be challenging to identify the sources of contaminants. This article studies the deployment of multi-cooperative unmanned aerial vehicles (UAVs) to look for sources of pollution in an unknown region. Specifically, a probabilistic search strategy based on the Local Particle Swarm Optimization (LoPSO) method is suggested to design an optimal strategy that enhances the cooperative search of drones while cutting down on overall search time and improving source detection efficiency. The entire strategy is divided into two phases: exploration and exploitation. In the first phase, the detection and tracing of the plume is the favored task, in which each UAV operates in either the Greedy or LoPSO mode and chooses its path based on plane coordinates generated according to the active mode. By utilizing the areas with a high probability of discovering the source on flight mode LoPSO, the search is focused during the exploitation phase on the precise search of the exact location of the pollutant source. This is done under the direction of a probabilistic computation that uses the Bayesian process model to create and update the probability map of the pollutant source location as new sensor data becomes available. The simulation data of the proposed technique demonstrates promising results in terms of the complexity and accuracy attained in identifying pollution sources.