Research On State Information Prediction Of UAV In Smart City

The next generation wireless network with Unmanned aerial vehicle(UAV)is considered to be one of the most effective solutions for improving the communication coverage.However,UAV is easily affected by wind and air pressure,accompanied by the certain position deviation and time delay during the air communication,so that the inaccurate beamforming will be performed by the ground base station(BS),causing unnecessary capacity loss.To address this issue,this paper is aimed at Internet of Things(Io T)communication system in smart city.Through Recurrent neural network,the next moment state information of UAV can be accurately predicted.The main work of this paper is as follows:First,this paper introduces the research status of UAV relay communication system,DOA estimation algorithm and Recurrent neural network,and gives the current elevation angle and horizontal angle of UAV based on the spectrum peak of DOA estimation algorithm of cylindrical antenna array.However,the existing DOA estimation algorithm only considers the current state information and does not consider the state information of next moment.Therefore,the performance of the communication transmission scheme based on DOA estimation algorithm will be degraded considering the position deviation and time delay.Then,to address the issue we discussed above,this paper considers to apply Recurrent neural network to predict the next moment state information of UAV in smart city.By analyzing the structure and characteristics of UAV angle data,a state information prediction model based on Recurrent neural network is proposed.According to the spatial spectrum of cylindrical antenna array’s DOA estimation algorithm above,the elevation angle and horizontal angle of current UAV can be obtained,thereby providing a reliable input dataset for the establishment of the prediction model.Through the training of the prediction model,the precise position information of next moment UAV can be predicted,so that the accurate beamforming can be performed by ground BS,thereby reducing the capacity loss of entire communication system.Finally,this paper builds a simulation platform to simulate city and highway sceneries under the smart Io T,analyzes and compares the accuracy of the prediction model under different parameter frameworks,and obtains the corresponding parameters with the highest accuracy.The simulation results show that the Recurrent neural network-based UAV state information prediction model proposed in this paper has high accuracy,and the comparison experiments show that the prediction effect and performance are higher than other prediction models,thereby improving the communication transmission of the system,reducing the capacity loss of entire communication system.