The Method of Rank Evaluation in Channel Estimation Problem Based on a Priori Data

Rank estimation is a complex and important task in modern communication systems. A number of operations can be significantly simplified by using rank as a stopping criterion, such as singular value decomposition, Tucker decomposition, etc. Singular value decomposition has a high computational complexity, but if the rank of the matrix is much lower than its dimension, knowing the rank can significantly reduce the computational complexity. Knowing the rank of the channel matrix allows for optimal use of the frequency-time resource to increase the bandwidth. Due to the presence of interference in the receiving and transmitting path, estimation accuracy may be reduced. This article discusses algorithms for estimating matrix ranks, presenting a new statistical method for estimating the rank, compares their effectiveness, and evaluates computational complexity.