Research On Multi-antenna Channel Measurement Feedback And Signal Detection Algorithm Based On Deep Learning

At present,5G is in full swing in China,and its status has been elevated to the level of national strategy.The mobile Internet brought by 4G has greatly facilitated people's lives and brought extremely high economic benefits.With the completion of 5G construction,it will definitely bring significant changes to people's lives,and the economic benefits brought by it will be immeasurable.Compared with the 4G era,5G emphasizes three aspects:high speed,low latency,and high reliability.These features also face the Internet of Everything and cloud computing scenarios.In order to meet the requirements of 5G communication,Massive MIMO has become the core technology in 5G communication.For Massive MIMO systems,if you need to maximize spectral efficiency and perform efficient,fast,and reliable data transmission,you need to obtain channel state information on the one hand,and efficient and accurate signal detection algorithms on the other.This article focuses on these two Research.For the measurement of channel state information,this article focuses on the performance improvement of the algorithm for channel state information compression feedback recovery in FDD scenarios.Based on a comparative study of traditional compressed sensing algorithms and deep learning-based approaches,this paper introduces attention mechanisms and recurrent neural network-based encoder-decoder structures in neural network algorithms,which greatly improves the performance of the algorithm.The simulation results show that under the approximate algorithm complexity,the performance of the proposed algorithm is much better than the performance of current algorithms based on deep learning and compressed sensing.For the direction of signal detection algorithms,in the Massive MIMO scenario,as the number of antennas increases,the complexity of similar ball decoding and maximum likelihood decoding will increase exponentially,and efficient detection algorithms have become a popular research area for scholars direction.This paper analyzes and studies the MIMO detection algorithm based on deep learning,introduces the attention mechanism,and further improves the performance of the algorithm.In summary,this paper combines the attention mechanism and network structure design in deep learning algorithms,and proposes novel algorithms in both channel measurement and signal detection to improve the performance of the algorithm in corresponding scenarios.