Energy-efficient UAV-BS-coordinated Data Collection for Wireless Sensor Networks of High-speed Railways

The wireless sensor based railway monitoring system faces challenges such as limited battery lifetime of sensor nodes (SNs). Moreover, information freshness of sensing data is a crucial metric for environment monitoring. In this paper, to extend SNs' battery lifetime and enhance information freshness of sensing data, we propose to employ unmanned aerial vehicles (UAVs), whose trajectory is planable, to cooperate with base station (BS) for data collection from wirelss sensor networks (WSNs). Then, we formulate the optimization problem with the objective to minimize the energy consumption of SNs and the age of information (AoI) which is used to evaluate the information freshness, by jointly adjusting the task allocations of data collection and the UAV trajectories. We find that our formulated optimization problem is a Markov decision process (MDP), based on which we propose a deep reinforcement learning (DRL)-based UAV-BS-coordinated data collection algorithm to find an asymptotically optimal solution. Simulation results demonstrate that our proposed DRL-based UAV-BS-coordinated data collection algorithm can significantly reduce AoI of sensing data and effectively extend the battery lifetime of the SNs compared to other baseline algorithms.