Deep learning multiple-input-multiple-output detection method

Abstract

As the essence of detection is division of sample space, a large amount of unnecessary computing resources are consumed. The invention provides a deep learning multiple-input-multiple-output detectionmethod, and the method comprises the steps: designing a deep learning model with low complexity and low bit error rate for low-order multiple-input-multiple-output; randomly generating a channel matrix, and generating a training set for training a deep learning model by using the channel matrix; generating a test set for testing the performance of the deep learning model under the time-invariantchannel by utilizing the channel matrix; adding a random channel matrix increment on the basis of the channel matrix to generate a training set and a test set; comparing with other deep learning models and a traditional multiple-input-output detection algorithm in the aspect of bit error rate; and compared with other deep learning models and a traditional multiple-input-output detection algorithmin the aspect of complexity. The detection algorithm is low in complexity, low in bit error rate and suitable for being applied to medium-low-cost communication equipment.

Description

technical field [0001] The present application belongs to the technical field of anomaly detection, and in particular relates to a deep learning multiple-input multiple-output detection method. Background technique [0002] Mobile communication has experienced development in recent decades, and wireless communication technology has penetrated into every aspect of life and has begun to truly change people's lives. The communication system continues to develop, and has developed from the first generation mobile communication system to the fourth generation mobile communication system. With the rapid development of the Internet, the number of terminals and the needs of users have increased greatly. The existing fourth-generation mobile communication system can no longer meet the needs of people in the future. Under this background, people have developed the fifth-generation mobile communication system. system. The deployment of 5G network has already started. The fifth-gener...

AI Technical Summary

Problems solved by technology

However, these networks have some shortcomings. Some networks are too complex, such as OAMPNet, which needs to invert a large number of matrices; some networks have poor detection performance under low-order MIMO, such as DetNet and MMNet, even if they have high-order MIMO. good detection performance

More importantly, the currently proposed networks designed for MIMO detection are not interpretable, and the specific functions of each layer of the network cannot be clearly known. Some of them introduce some parameters into the original detection algorithm, but there is no clear description of the introduction of these The meaning of parameters, and why the introduction of these parameters can improve network performance, such as OAMPNet, MMNet; some use neural networks to solve a non-convex problem, but the iterative formula used is theoretically infeasible, such as DetNet, ScNet

In essence, the compressed storage structure of these networks is too complex and unexplainable

Images