Block Diagonal Sparse Bayesian Channel Estimation Method for Sc-mimo Underwater Acoustic Communication Environment

Abstract

The invention discloses a block diagonal sparse Bayesian channel estimation method in an SC-MIMO underwater acoustic communication environment. The method is suitable for an underwater acoustic channel which is rapid in time varying, obvious in Doppler effect and serious in multipath interference. A blocking sparse Bayesian algorithm provided by the invention is used for channel estimation, and ischaracterized in that the spatial correlation and sparsity of an MIMO underwater acoustic channel and the statistical property of the channel are fully utilized; a channel block diagonal model is constructed, the spatial correlation of a corresponding channel by each sub-block is described, and an iterative estimation channel coefficient, a covariance and a noise parameter are jointly updated bycombining expectation maximization algorithm. Compared with a traditional Bayesian learning algorithm, the method is high in algorithm robustness and calculation complexity, and is high in estimationprecision than an OMP algorithm and an IPNLMS algorithm.

Description

technical field [0001] The invention belongs to the field of underwater acoustic communication, and relates to a block diagonal sparse Bayesian channel estimation method, which is suitable for underwater channel environments with rapid time-varying and severe multipath interference. Background technique [0002] There are great technical challenges in the field of underwater communication, which are mainly reflected in three aspects: multi-path delay leads to serious inter-symbol interference and the interference lasts for a long time; Puller effect; the very limited bandwidth of underwater channels limits the transmission rate of underwater communications. In order to realize underwater high-speed data transmission, the space-time diversity gain brought by MIMO technology makes it more and more widely used in the field of underwater high-speed communication. However, it also faces some technical difficulties. The strong spatial correlation inevitably reduces the gain broug...

AI Technical Summary

Problems solved by technology

A typical example is the LMS least mean square algorithm, but the channel impulse response is required to remain constant within the duration of a data block. However, the constraints are easily broken by the rapidly time-varying dynamic environment underwater

Another method is to estimate the Doppler shift, but it brings high computational complexity

Recently, the sparse Bayesian learning algorithm has been applied to underwater acoustic channel estimation, which can better deal with the problem of channel over-parameterization, but it will also bring huge computational complexity.

At the same time, the channel covariance matrix is ​​assumed to be a diagonal matrix, which is not suitable in the underwater acoustic communication environment

Images