Channel estimation algorithm combining corrected angle mismatch with sparse Bayesian learning

Abstract

The invention discloses a channel estimation method combining corrected angle mismatch with sparse Bayesian learning, belongs to the technical field of wireless communication, the invention particularly relates to a channel estimation problem of an uplink angle domain of a large-scale multiple-input-multiple-output system. The method specifically comprises the following steps: a channel is preliminarily estimated by utilizing an orthogonal pilot frequency sequence sent by multiple users; the problem of angle mismatch is considered in the next step of accurate estimation; namely, in a virtual angle domain, the real angle of arrival does not necessarily fall on a pre-divided virtual angle; therefore, the angle deviation is used as an unknown parameter; wherein the linear fitting modeling ofthe angle deviation adopts a linear interpolation model and a first-order Taylor expansion model respectively, the angle deviation and a sparse gain coefficient are used as combined unknown variables,and under the framework of a sparse Bayesian learning algorithm, deviation information is corrected by continuously iterating and updating parameters, so that the sparse information of the channel isaccurately positioned.

Description

technical field [0001] The invention belongs to the technical field of wireless communication, in particular to a channel estimation algorithm combining corrected angle mismatch and sparse Bayesian learning. Background technique [0002] Sparse channel estimation has always been a research hotspot among scholars, but how to effectively obtain uplink channel state information in the angle domain is rarely studied. In fact, after discrete Fourier transform (Discrete FourierTransform, DFT), the channel has hidden sparsity in the virtual angle domain. The channel estimation problem in the angle domain usually uses the hidden sparsity in the virtual angle domain for channel estimation, the so-called virtual angle The domain refers to the division of a certain range of angle intervals into multiple uniform and discrete preset angles (sampling grid). However, the rough division of the preset angles will make the real angle of arrival (DOAs) deviate from the preset angle, resulting ...

AI Technical Summary

Problems solved by technology

[0002] The sparse channel estimation problem has always been a research hotspot of scholars, but how to effectively obtain the uplink channel state information in the angle domain is rarely studied.

However, the rough division of preset angles will cause the real angle of arrival (DOAs) to deviate from the preset angle, resulting in channel energy leakage; the fine division of sampling preset angles will not only increase the computational complexity but also increase the correlation within the measurement matrix. Deviates from RIP characteristics

Images