Research On Channel Intelligent Access Technology In UAV Aided Power Consumption Information Acquisition Network

With the rapid development of the smart grid,the trend of the explosive growth of power transmission data has put forward higher requirements on the reliability of electricity consumption information collection networks.However,the ground base stations of the electricity information collection network are easily restricted by factors such as fixed location and high deployment cost,and the collection of electricity information in remote areas also faces great difficulties.For this reason,this paper uses the characteristics of UAVs,such as wide coverage and fast deployment,to build an aerial communication platform to assist the wireless communication of electricity consumption information collection networks.However,when assisting wireless communication,the uncertainty of UAV location and the limited energy make the communication performance greatly reduced,and the dynamic nature of the channel environment also directly affects the UAV multi-channel access process.Therefore,the optimal trajectory design and the best power allocation method can effectively improve the communication performance,and the intelligent access method of multi-channel can also enable the UAV to cope well with the complex and changing communication environment.In this paper,based on the investigation of the joint trajectory design of UAVs and the optimization problem of power allocation,we investigate the UAV access multichannel method to increase the success rate of electricity information collection and improve the communication quality of the electricity information collection network.The main contributions of this paper include the following two aspects.1.To maximize the communication performance and system capacity,this paper proposes a joint trajectory design and power allocation optimization algorithm based on deep reinforcement learning.First,a UAV system model is established,and the joint UAV trajectory design and power allocation optimization problem is developed considering constraints such as flight collision and quality of service.For the nonconvexity and combinatoriality of this optimization problem,the optimization problem is described as a Markov decision process.On this basis,considering the continuous action space characteristic of the Markov decision process,a deep deterministic policy gradient algorithm is proposed to combine the policy-based learning method and the value-based learning method to obtain the accurate state transfer probability.Simulation experimental results show that the algorithm proposed in this paper outperforms other optimization methods in terms of overall network utility performance and has a faster convergence rate.2.For the problem of dynamic change of channels in electricity information collection networks,this paper proposes a multi-channel intelligent access method based on deep reinforcement learning.First,by describing the multichannel access model as a Markov decision process,the Q-learning method is proposed to realize the intelligent access of multichannel.Based on this,a deep neural network is designed to obtain a nearoptimal multichannel intelligent access strategy for the characteristics of Q-learning such as large state space and slow convergence speed.Finally,the NS3 simulation platform is built to verify the performance of the proposed multi-channel intelligent access method.The simulation results show that the proposed deep reinforcement learning-based multichannel intelligent access method can achieve better access performance in dynamic multi-channel environments with faster convergence than existing reinforcement learning methods.