Energy-Efficient UAV Communications Under Stochastic Trajectory: A Markov Decision Process Approach

Abstract

A typical application of unmanned aerial vehicles (UAVs) is the surveillance of distant targets, where data collected by its cameras need to be transmitted back to a ground terminal (GT) for further processing. However, due to the high mobility of a UAV, the power consumption for transmitting data tasks might be very high, especially when the UAV is far away from the GT. To guarantee the energy-efficiency of the UAV, the collected data tasks can be preprocessed to reduce the amount of data transmitted, which can potentially save the power consumption of the UAV. In this article, a random walk model is implemented to capture the mobility of the UAV. Our design goal is to minimize the average power consumption of the UAV under the delay constraint. We adopt the probabilistic scheduling based on the Markov Decision Process (MDP) framework to schedule the transmission and computation of the data tasks in the UAV. The power-optimal scheduling policy can be obtained by converting the joint optimization problem to linear programming (LP), and thus the optimal power-delay tradeoff can be achieved. Finally, the optimization results are validated with extensive simulations.