NOMA-Aided UAV Data Collection from Time-Constrained IoT Devices

Abstract

Non-orthogonal multiple access (NOMA) is one of the promising access technologies to improve spectral efficiency and serve a higher number of users simultaneously. The latter proves important in time-sensitive services when data has to be collected before a set deadline, otherwise, the data is rendered useless. Therefore, in this paper, we utilize a NOMA-aided unmanned aerial vehicle (UAV) for data collection from time-constrained IoT devices. We optimize the trajectory of the UAV, IoT devices scheduling, and power allocation to maximize the number of served devices while considering the constraints of UAV energy and flight duration, and NOMA clustering. Given the complexity of the problem and the incomplete knowledge about the environment, it is divided into two subproblems. In the first subproblem, the UAV trajectory and the selection of the first device in the NOMA cluster at each time slot are modeled as a Markov Decision Process, and Proximal Policy Optimization is used to solve it. For the second device selection, a heuristic algorithm is used based on prioritizing devices with higher bit rate requirements and strict deadlines. The second subproblem considers power allocation inside the NOMA cluster, where it is formulated as an optimization problem for maximizing the sum rates of the two selected users. Finally, we demonstrate the performance gains of our solution in different scenarios while varying the system parameters as compared with alternative approaches. In particular, our proposed solution achieves a 10% to 30% performance gain compared to the traditional orthogonal multiple access scheme.