On the GA-based UAV-Assisted Data Collection with Stochastic Arrivals in Remote Sensing Applications
Minbin Li, Qingchun Chen · 2025
Unmanned Aerial Vehicle (UAV)-assisted wireless sensor networks (WSNs) have emerged as a promising paradigm for enabling reliable data acquisition from distributed sensor nodes (SNs) in geographically constrained environments, such as remote wilderness or post-disaster terrains. In large-scale heterogeneous WSNs with stochastic data arrivals, the UAV-assisted data collection design becomes crucial to optimize aerial resource utilization while maintaining reliable data collection. This paper investigates a multi-UAV collaborative data collection framework that holistically integrates ground SN geo-distributions and their stochastic data arrivals. To address this NP-hard optimization problem, we propose a metaheuristic deployment strategy leveraging an enhanced genetic algorithm (GA) with adaptive crossover-mutation operators, aiming to minimize UAV fleet size under strict constraints of all SNs coverage and reliable data collection task. Numerical results reveal that, the proposed GA-based UAV-assisted data collection scheme can determine reasonable multiple UAV deployment to guarantee full ground SNs coverage with the least required UAV number, outperforming greedy heuristic and geo-distribution clustering-based approaches. Furthermore, the proposed UAV-assisted data collection exhibits remarkable adaptability, maintaining service stability across diverse SN geo-distributions and stochastic arrival intensities. Analysis results indicate that, the required number of UAVs will be dependent on the number and geo-distribution of ground SNs. For the given ground geodistribution of all SNs, only when the data arrival rates exceed a specific threshold, the required number of UAVs gradually increases with rising data arrival rates.