Precoder Design in Multi-User FDD Systems with VQ-VAE and GNN
This work addresses efficient precoding for multi-user wireless systems, offering incremental improvements by reducing pilot or feedback bit requirements compared to conventional methods.
The paper tackles robust precoder design in multi-user FDD wireless systems by using a VQ-VAE to overcome scalability issues of previous GMM-based methods, achieving significant sum rate gains through an end-to-end model that includes joint training with GNN and pilot optimization.
Robust precoding is efficiently feasible in frequency division duplex (FDD) systems by incorporating the learnt statistics of the propagation environment through a generative model. We build on previous work that successfully designed site-specific precoders based on a combination of Gaussian mixture models (GMMs) and graph neural networks (GNNs). In this paper, by utilizing a vector quantized-variational autoencoder (VQ-VAE), we circumvent one of the key drawbacks of GMMs, i.e., the number of GMM components scales exponentially to the feedback bits. In addition, the deep learning architecture of the VQ-VAE allows us to jointly train the GNN together with VQ-VAE along with pilot optimization forming an end-to-end (E2E) model, resulting in considerable performance gains in sum rate for multi-user wireless systems. Simulations demonstrate the superiority of the proposed frameworks over the conventional methods involving the sub-discrete Fourier transform (DFT) pilot matrix and iterative precoder algorithms enabling the deployment of systems characterized by fewer pilots or feedback bits.