UAV-Based Connectivity Maintenance for Borderline Detection

The communication aspects of Swarm-based Unmanned Aerial Systems (UAS) for surveillance and monitoring tasks have received increased attention in recent research. In this paper, we address the specific challenges of unpredictable incidents as forest fires, accidents in chemical or nuclear plants which cause potentially dangerous plumes. Therefore, we focus on the detection of chemical plume borderlines while maintaining the connectivity within the UAV team. Leveraging previously developed spatial exploration algorithms like Cooperative Repelling Walk we introduce two novel strategies to optimize the trade-off between detection-efficiency and communication constraints: the Aerosol-detecting Cooperative Repelling Walk (ADCRW) and the Distributed Dispersion Detection (DDD) Algorithm. Both algorithms are evaluated with the help of a multi-scale simulation model which includes a newly incorporated Briggs plume model. Considering the mentioned scenario the results of the performance evaluation demonstrate that the detection-efficiency of our novel algorithms is significantly improved, while at the same time connectivity goals are still met: with the new algorithms the time to detect the borderline of a plume can be reduced by 35%, while the connectivity within the swarm is about 100%.