Sparse Bayesian Multi-Task Learning of Time-Varying Massive MIMO\n Channels with Dynamic Filtering

2019/11/18 by Arash Shahmansoori, Shahmansoori, Arash
Computer Science · Engineering · #Blind Source Separation Techniques #Advanced MIMO Systems Optimization #Advanced Adaptive Filtering Techniques

paper · pdf · doi:10.48550/arxiv.1911.07570

Abstract

Sparsity of channel in the next generation of wireless communication for\nmassive multiple-input-multiple-output (MIMO) systems can be exploited to\nreduce the overhead in the training. The multitask (MT)-sparse Bayesian\nlearning (SBL) is applied for learning time-varying sparse channels in the\nuplink for multi-user massive MIMO orthogonal frequency division multiplexing\nsystems. In particular, the dynamic information of the sparse channel is used\nto initialize the hyperparameters in the MT-SBL procedure for the next time\nstep. Then, the expectation maximization based updates are applied to estimate\nthe underlying parameters for different subcarriers. Through the simulation\nstudies, it is observed that using the dynamic information from the previous\ntime step considerably reduces the complexity and the required time for the\nconvergence of MT-SBL algorithm with negligible sacrificing of the estimation\naccuracy. Finally, the power leakage is reduced due to considering angular\nrefinement in the proposed algorithm.\n

Related

• Enhanced Sparse Bayesian Learning Methods with Application to Massive MIMO Channel Estimation
• Time-Varying Downlink Channel Tracking for Quantized Massive MIMO Networks
• Variational Bayesian Inference for Time-Varying Massive MIMO Channels: Estimation and Detection
• Low-Complexity Message Passing Based Massive MIMO Channel Estimation by Exploiting Unknown Sparse Common Support with Dirichlet Process
• Nonlinear Sparse Bayesian Learning Methods with Application to Massive MIMO Channel Estimation with Hardware Impairments