A deep learning DOA estimation method based on original IQ data

By using the original IQ data and a convolutional classification network with an adaptive number of snapshots, features are directly extracted from the original data for DOA estimation, which solves the problem of information loss in traditional methods and achieves better DOA estimation performance and scene adaptability.

CN116840776BActive Publication Date: 2026-06-30HANGZHOU DIANZI UNIV +1
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
HANGZHOU DIANZI UNIV
Filing Date
2023-07-07
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

Traditional DOA estimation methods require a large number of snapshots to accurately estimate the direction of arrival (DOA), and converting the raw data into a covariance matrix leads to information loss, affecting the feature extraction and learning performance of deep neural networks.

Method used

Using raw IQ data as input to a deep neural network, an adaptive snapshot number convolutional classifier network is designed to learn and extract features directly from the raw data. A ResNet-structured convolutional classifier network is then constructed for DOA estimation.

Benefits of technology

It improves DOA estimation performance, maintains good performance under different signal-to-noise ratios and noise scenarios, has strong scene generalization ability, and adapts to input signals with different snapshot numbers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116840776B_ABST
    Figure CN116840776B_ABST
Patent Text Reader

Abstract

The application discloses a deep learning DOA estimation method based on original IQ data. The application uses I and Q components of the original signal as the input of the model to improve the performance. The application aims to solve the DOA estimation problem of a single signal source, and models the single signal source DOA estimation problem as a single-label multi-classification problem. By discretizing the DOA range, the possible directions of arrival are taken as corresponding labels. A convolutional neural network is designed to adapt to different numbers of snapshots, and accurate DOA estimation can be adaptively obtained for input signals of different lengths. Experimental results show that, compared with existing deep learning DOA estimation methods based on covariance matrix as input, the scheme has more excellent performance, and can provide a more reliable solution for array signal processing.
Need to check novelty before this filing date? Find Prior Art