Energy-Efficient Deployment and Power allocation for VLC-enabled Unmanned Aerial Vehicles Communications

Deploying unmanned aerial vehicles (UAV) enabled visible light communication (VLC) networks to accommodate both the illumination and communication requirements of all users presents a serious challenge. In this paper, the UAV-VLC network deployment is investigated by the joint optimization of user association, UAV placement and power allocation, which is mathematically formulated as a minimization of energy consumption problem. The original problem is decoupled into two subproblems and sequentially solved by the proposed two-stage optimization scheme. To elaborate, the K-means algorithm is employed to cluster the users firstly, thereby establishing the user association indirectly. Then, the deep reinforcement learning (DRL) based technique is employed to determine the optimum UAV placement and power allocation as considering the specific requirements of illumination and communications for all users. Simulation results demonstrate that the proposed scheme can achieve the superior performance and can reduce the total transmit power consumption at least by 74.39% and 67.62% as compared with the conventional schemes.