15 February 2022 in IEEE Transactions on Wireless Communications
We propose an innovative machine learning-based technique to address the problem of channel acquisition at the base station in frequency division duplex systems. In this context, the base station reconstructs the full channel state information in the downlink frequency range based on limited downlink channel state information feedback from the mobile terminal. The channel state information recovery is based on a convolutional neural network which is trained exclusively on collected channel state samples acquired in the uplink frequency domain. No acquisition of training samples in the downlink frequency range is required at all. Finally, after a detailed presentation and analysis of the proposed technique and its performance, the “reciprocity of distribution” assumption of the convolutional neural network that is central to the proposed approach is validated with an analysis based on the maximum mean discrepancy metric.
25 October 2022 in IEEE Transactions on Wireless Communications
To fully unlock the benefits of multiple-input multiple-output (MIMO) networks, downlink channel state information (CSI) is required at the base station (BS). In frequency division duplex (FDD) systems, the CSI is acquired through a feedback signal from the user equipment (UE). However, this may lead to an important overhead in FDD massive MIMO systems. Focusing on these systems, in this study, we propose a novel strategy to design the CSI feedback. Our strategy allows to optimally design variable length feedback, that is promising compared to fixed feedback since users experience channel matrices differently sparse. Specifically, principal component analysis (PCA) is used to compress the channel into a latent space with adaptive dimensionality. To quantize this compressed channel, the feedback bits are smartly allocated to the latent space dimensions by minimizing the normalized mean squared error (NMSE) distortion. Finally, the quantization codebook is determined with k -means clustering. Numerical simulations show that our strategy improves the zero-forcing beamforming sum rate by 17%, compared to CsiNetPro. The number of model parameters is reduced by 23.4 times, thus causing a significantly smaller offloading overhead. At the same time, PCA is characterized by a lightweight unsupervised training, requiring eight times fewer training samples than CsiNetPro.
4 August 2023 in IEEE Wireless Communications Letters
We propose a user-driven method for adaptive quantization and feedback of channel state information (CSI) in frequency division duplexing systems. The approach can be combined with existing autoencoder neural networks and requires only a single codebook shared between the users and the base station. Users can provide feedback with different numbers of bits depending on their data rates, without system changes or additional storage requirements. We optimize CSI feedback by combining the nested dropout layer with vector quantization, resulting in a variable-length feedback approach. Simulation results show the proposed method outperforms state-of-the-art schemes in dynamically adapting to different feedback rates.
18 December 2023 in IEEE Open Journal of Signal Processing
Channel state information (CSI) is crucial for enhancing the performance of wireless systems by allowing to adjust the transmission strategies based on the current channel conditions. However, obtaining precise CSI is difficult because of the fast-changing channel conditions caused by multi-path fading. An inaccurate CSI hinders the performance of various adaptive wireless systems, highlighting the need for channel prediction techniques to effectively mitigate the drawbacks of outdated CSI. Conventional methods typically depend on assumptions regarding user velocity or require knowledge of the Doppler frequency. In contrast to existing approaches, we aim for a more robust and practical solution by training neural networks without making any assumptions about user velocity, relying solely on noisy channel observations during training. Specifically, we adapt both the sequence-to-sequence with attention (Seq2Seq-attn) and transformer models for channel prediction. Additionally, a new technique called reverse positional encoding is introduced in the transformer model to improve the robustness of the model against varying sequence lengths. Similarly, the encoder outputs of the Seq2Seq-attn model are reversed prior to the application of attention mechanisms. By means of simulations, we show that these proposed techniques enable the models to effectively capture relationships within sequences of channel snapshots without increasing the complexity. Importantly, this capability remains robust across varying sequence lengths, representing a substantial improvement over existing methodologies.