A low-complexity algorithm based on variational Bayesian inference for MIMO channel estimation

•The complexity of the MIMO-SBL algorithm for channel estimation increases cubically with the number of transducers.•We propose the MIMO-VBI algorithm by utilizing variational Bayesian inference to separate the update processes for individual channels and approximate their posteriors.•The complexity of MIMO-VBI increases linearly with the number of transducers, and it achieves better estimation performance compared to MIMO-SBL.