Salvato in:
| Autori principali: | , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2022
|
| Soggetti: | |
| Accesso online: | https://arxiv.org/abs/2207.06924 |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866918135327621120 |
|---|---|
| author | Rizzello, Valentina Nerini, Matteo Joham, Michael Clerckx, Bruno Utschick, Wolfgang |
| author_facet | Rizzello, Valentina Nerini, Matteo Joham, Michael Clerckx, Bruno Utschick, Wolfgang |
| contents | In this work, we propose an efficient method for channel state information (CSI) adaptive quantization and feedback in frequency division duplexing (FDD) systems. Existing works mainly focus on the implementation of autoencoder (AE) neural networks (NNs) for CSI compression, and consider straightforward quantization methods, e.g., uniform quantization, which are generally not optimal. With this strategy, it is hard to achieve a low reconstruction error, especially, when the available number of bits reserved for the latent space quantization is small. To address this issue, we recommend two different methods: one based on a post training quantization and the second one in which the codebook is found during the training of the AE. Both strategies achieve better reconstruction accuracy compared to standard quantization techniques. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2207_06924 |
| institution | arXiv |
| publishDate | 2022 |
| record_format | arxiv |
| spellingShingle | Learning Representations for CSI Adaptive Quantization and Feedback Rizzello, Valentina Nerini, Matteo Joham, Michael Clerckx, Bruno Utschick, Wolfgang Signal Processing Machine Learning In this work, we propose an efficient method for channel state information (CSI) adaptive quantization and feedback in frequency division duplexing (FDD) systems. Existing works mainly focus on the implementation of autoencoder (AE) neural networks (NNs) for CSI compression, and consider straightforward quantization methods, e.g., uniform quantization, which are generally not optimal. With this strategy, it is hard to achieve a low reconstruction error, especially, when the available number of bits reserved for the latent space quantization is small. To address this issue, we recommend two different methods: one based on a post training quantization and the second one in which the codebook is found during the training of the AE. Both strategies achieve better reconstruction accuracy compared to standard quantization techniques. |
| title | Learning Representations for CSI Adaptive Quantization and Feedback |
| topic | Signal Processing Machine Learning |
| url | https://arxiv.org/abs/2207.06924 |