In the near-field region of an extremely large-scale MIMO (XL MIMO) system, channel reconstruction can be solved by utilizing sparse parameter estimation after transforming the received pilots at the base station (BS) into Cartesian domain. However, the process of exhaustive search over the codebook consumes significant computational resources and running time, especially when dealing with a vast number of antennas. In this study, we visualize the sparse channel matrix in the Cartesian domain as an channel image and propose a deep neural network, i.e., channel keypoint detection network (CKNet), to locate the user and scatterers. Subsequently, we employ a straightforward Newton optimization module to fine-tune the estimations. Experimental results demonstrate that the CKNet-empowered channel reconstruction scheme substantially reduces computational complexity while maintaining high accuracy in both user and scatterer localization and channel reconstruction.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Efficient Near-Field User Localization and Channel Reconstruction via Image Keypoint Detection


    Contributors:
    Li, Mengyuan (author) / Han, Yu (author) / Jin, Shi (author)


    Publication date :

    2024-06-24


    Size :

    2124604 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Keypoint based action localization

    KADAV ASIM / LAI FARLEY / GRAF HANS PETER et al. | European Patent Office | 2025

    Free access

    KEYPOINT BASED ACTION LOCALIZATION

    KADAV ASIM / LAI FARLEY / GRAF HANS PETER et al. | European Patent Office | 2022

    Free access

    Progressive Keypoint Detection With Dense Siamese Network for SAR Image Registration

    Xiang, Deliang / Xu, Yihao / Cheng, Jianda et al. | IEEE | 2023


    MONOCULAR 2D SEMANTIC KEYPOINT DETECTION AND TRACKING

    CHEN HAOFENG / ARJUN BHARGAVA / RARES ANDREI AMBRUS et al. | European Patent Office | 2023

    Free access