Deep Learning Signal Detection for Large-Scale MIMO
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The complexity of signal detection in large-scale millimeter wave MIMO systems increases due to the larger number of transceiver antennas, leading to higher computational complexity and performance degradation compared to traditional MIMO systems.
Innovation Solution
A signal detection method that involves obtaining a channel matrix through channel estimation, performing data preprocessing such as QR decomposition, and inputting the resulting matrices and vectors to a trained deep learning model comprising N convolutional network models, where N equals the number of transmission antennas, to obtain an estimated detection signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of transceiver antennas is increased to improve MIMO performance, then signal detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent divides the signal detection process into multiple stages: channel estimation, data preprocessing (QR decomposition), and deep learning-based detection. This segmentation allows each component to be optimized independently, reducing overall computational complexity while maintaining detection accuracy in large-scale MIMO systems
Solution Approach 2:
The patent introduces a trained deep learning model as an intermediary between the preprocessed channel data and the final detection output. This intermediary learns optimal detection strategies during training, enabling accurate signal detection with reduced computational complexity during actual operation compared to traditional maximum likelihood detection
2Reliability
If traditional MIMO detection methods are used in large-scale systems, then detection performance is maintained, but processing time increases
Solution Approach 1:
The patent performs channel estimation and QR decomposition preprocessing before the actual detection process. The deep learning model is trained in advance on representative channel conditions, so during operation, detection can be performed rapidly using the pre-trained model without repeating complex computations, thereby reducing processing time while maintaining performance
Data Source
AI summary
The present disclosure provides signal detection methods and apparatuses. A method executed by a network side device includes: obtaining a channel matrix by performing a channel estimation on a signal received from a terminal device; performing a data preprocessing on the channel matrix to obtain a first matrix and a first vector; and inputting the first matrix and the first vector into a trained deep learning model to obtain an estimated detection signal, wherein the trained deep learning model includes N trained convolutional network models, and N is a positive integer equal to a number of transmitting antennas.