A Research On Anti-jamming Communication Transmission System Based On Intelligent Beamforming

Millimeter wave has abundant spectrum resources,which can greatly alleviate the pressure of spectrum resources and meet the requirements of future 5G multimedia services such as rate and delay.Therefore,it has become the preferred physical layer technology in the 5G NR specification.However,due to the strong fading characteristics of millimeter waves,the advantages of its large bandwidth need to be demonstrated with the help of beamforming architecture,and the millimeter wave system based on the digital-analog hybrid structure will be limited by the hardware constraints of the phase shifter and the transmission power of the antenna.This makes the objective function of the hybrid beamforming algorithm a non-convex optimization problem,which is difficult to solve by traditional optimization theory.In addition,the performance of the hybrid beamforming algorithm will also be affected by incomplete channel state information,imperfect hardware,etc.,the environment adaptability is poor,and it is difficult to maintain good performance in multiple communication scenarios.In particular,due to the sparse nature of the millimeter wave channel,the available paths contained in the channel are few.If the transmission encounters interference,it is also a major challenge to its antiinterference ability.In this paper,the anti-interference transmission of large-scale MIMO millimeter wave system is taken as the research content,and the realization of low-complexity and high-performance intelligent beamforming algorithm as the research entry point,and a hybrid beamforming algorithm based on neural network is proposed.The proposed algorithm can independently select orthogonal transmission and anti-interference transmission in under-rank channels according to the characteristics of the channel,which solves the non-convex optimization problem in the design of beamforming algorithms and achieves better performance than traditional beamforming algorithms.In addition,this paper proposes that the learning rate adaptive algorithm can effectively improve the efficiency of neural network training and shorten the training time for the problem of long-term neural network training iteration.Finally,the algorithm is proved from incomplete channels and low-resolution shifters.Robustness is of great significance for improving the performance of hybrid beamforming systems.Compared with the traditional hybrid beamforming algorithm,the neural networkbased hybrid beamforming precoding algorithm proposed in this paper has the following innovations:1 The beamforming architecture replaces the beamforming matrix with a neural network.When the channel is not overloaded(that is,the number of data streams is less than the rank of the channel matrix),the orthogonal transmission mode is used.On the under-rank channel(that is,the data stream is larger than the channel matrix)When transmitting under rank),it can effectively reduce the interference between multiple data streams,realize anti-interference under-rank transmission,and can provide better bit error rate performance than the traditional hybrid beamforming algorithm.2 The beamforming architecture can effectively resist errors caused by channel estimation and interference caused by non-ideal hardware.Since the neural network is not sensitive to small errors in the data,the architecture uses the data fitting ability of the neural network to compensate for the error of channel estimation and the defects caused by non-ideal hardware,so that the system has strong anti-interference ability and robustness.Greatness.3 The architecture can effectively avoid the problems of neural network training requiring a large number of iterations and too long training time based on the adaptive learning rate algorithm,which realizes the automatic adjustment of the optimal learning rate of the neural network during the training process,which can effectively improve the neural network The training efficiency of the system,while shortening the training time of the neural network,can still provide better performance than the traditional beamforming algorithm.This allows the architecture to match channel changes and still maintain certain performance advantages under changing channels.In summary,the neural network-based beamforming precoding algorithm proposed in this paper can not only perform orthogonal transmission when the channel is not overloaded,but also reduce the interference between the data streams under the underrank channel and realize the channel under-rank transmission.In addition,the architecture can also effectively resist interference in various communication scenarios such as channel estimation errors and non-ideal hardware.Therefore,the architecture can use neural networks to resist various interferences in communication scenarios and maintain certain communication performance advantages,so it has strong anti-interference and flexible reliability.