UAV-Enabled Data Collection for IoT Networks via Rainbow Learning

TL;DR

This paper introduces a novel deep reinforcement learning approach for optimizing UAV trajectories and resource allocation to maximize data collection efficiency in IoT networks.

Contribution

It proposes a double-loop DRL algorithm, rainbow learning, for joint optimization of UAV trajectory, scheduling, and power allocation, addressing a complex non-convex problem.

Findings

Proposed algorithms outperform benchmarks in data collection volume.

Significant improvements in energy efficiency and fairness.

Effective joint optimization enhances UAV-enabled IoT data collection.

Abstract

Unmanned aerial vehicles (UAVs) enabled Internet of things (IoT) systems have become an important part of future wireless communications. To achieve higher communication rate, the joint design of UAV trajectory and resource allocation is crucial. In this paper, a multi-antenna UAV is dispatched to simultaneously collect data from multiple ground IoT nodes (GNs) within a time interval. To improve the sum data collection (SDC) volume from the GNs, the UAV trajectory, the UAV receive beamforming, the scheduling of the GNs, and the transmit power of the GNs are jointly optimized. Since the problem is non-convex and the variables are highly coupled, it is hard to be solved using traditional methods. To find a near-optimal solution, a double-loop structured optimization-driven deep reinforcement learning (DRL) algorithm, called rainbow learning based algorithm (RLA), and a fully DRL-based algorithm are proposed to solve the problem effectively. Specifically, the outer-loop of the RLA utilizes a fusion deep Q-network to optimize the UAV trajectory, GN scheduling, and power allocation, while the inner-loop optimizes receive beamforming by successive convex approximation. Simulation results verify that the proposed algorithms outperform two benchmarks with significant improvement in SDC volumes, energy efficiency, and fairness.