A joint self-information and Markovian model-driven deep channel estimation and feedback model for millimeter-wave massive MIMO systems.
Saved in:
| Title: | A joint self-information and Markovian model-driven deep channel estimation and feedback model for millimeter-wave massive MIMO systems. |
|---|---|
| Authors: | Kadiyala, Ragodaya Deepthi1 (AUTHOR) kr720059@student.nitw.ac.in, Lokam, Anjaneyulu1 (AUTHOR), Bhar, Chayan1 (AUTHOR) |
| Source: | Wireless Networks (10220038). Feb2025, Vol. 31 Issue 2, p1683-1703. 21p. |
| Subjects: | Image processing equipment, Bit error rate, Information & communication technologies, MIMO systems, Wireless communications, Channel estimation |
| Abstract: | The potential benefits of intelligent wireless communications with millimeter and massive multiple-input multiple-output (M-MIMO) are depend on the obtainability of channel state information (CSI) at the base station (BS). But, in frequency division duplex systems, the absence of channel reciprocity does not contribute to the challenge of attaining precise CSI. To address this issue, numerous researchers investigated efficient deep learning-based architectures and revealed the efficacy particularly for channel recovery and compression. Nonetheless, the existing methods suffers from higher complexity, less achievable rate, and performance degradation. Aiming at these problems, in this paper, a novel joint Self-information and Markovian model-driven Deep Channel Estimation and Feedback (JSM-DCEF) model is proposed for millimeter M-MIMO system. Initially, a self-information adjustment module is developed to pre-compress the raw CSI image for removing redundancy and determining self-information based on the structural features. Then, a residual wide-kernel deep convolutional auto-encoder network (RWK-DCAE) is used to encode CSI at User Equipment (UE). Similarly, to decode CSI, a residual wide-kernel deep convolutional auto-decoder network (RWK-DCAD) is used BS. The RWK-DCAE consists of wide-kernel convolutional layer and residual connections to learn the CSI features effectively. Additionally, Markovnet is coupled with proposed encoder and decoder architecture to differentially decode the feedback CSI. Also, this model is used to increase the bandwidth effectiveness and reconstruction accuracy by providing CSI data of two consecutive time slots as input for encoder. The simulation results are analyzed in terms of bit error rate (BER), mean square error (MSE), and spectral efficiency with respect to different SNR levels and training epochs. As a result, the numerical outcomes proves that the JSM-DCEF model outperforms the existing models and attains higher spectral efficiency of 20 bps/Hz at 10 Db SNR, respectively. [ABSTRACT FROM AUTHOR] |
| Copyright of Wireless Networks (10220038) is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Engineering Source |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | The potential benefits of intelligent wireless communications with millimeter and massive multiple-input multiple-output (M-MIMO) are depend on the obtainability of channel state information (CSI) at the base station (BS). But, in frequency division duplex systems, the absence of channel reciprocity does not contribute to the challenge of attaining precise CSI. To address this issue, numerous researchers investigated efficient deep learning-based architectures and revealed the efficacy particularly for channel recovery and compression. Nonetheless, the existing methods suffers from higher complexity, less achievable rate, and performance degradation. Aiming at these problems, in this paper, a novel joint Self-information and Markovian model-driven Deep Channel Estimation and Feedback (JSM-DCEF) model is proposed for millimeter M-MIMO system. Initially, a self-information adjustment module is developed to pre-compress the raw CSI image for removing redundancy and determining self-information based on the structural features. Then, a residual wide-kernel deep convolutional auto-encoder network (RWK-DCAE) is used to encode CSI at User Equipment (UE). Similarly, to decode CSI, a residual wide-kernel deep convolutional auto-decoder network (RWK-DCAD) is used BS. The RWK-DCAE consists of wide-kernel convolutional layer and residual connections to learn the CSI features effectively. Additionally, Markovnet is coupled with proposed encoder and decoder architecture to differentially decode the feedback CSI. Also, this model is used to increase the bandwidth effectiveness and reconstruction accuracy by providing CSI data of two consecutive time slots as input for encoder. The simulation results are analyzed in terms of bit error rate (BER), mean square error (MSE), and spectral efficiency with respect to different SNR levels and training epochs. As a result, the numerical outcomes proves that the JSM-DCEF model outperforms the existing models and attains higher spectral efficiency of 20 bps/Hz at 10 Db SNR, respectively. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 10220038 |
| DOI: | 10.1007/s11276-024-03845-8 |
