Parallel enhanced CLMS adaptive filtering method for processing non-circular signals

Abstract

The invention provides a parallel enhanced CLMS adaptive filtering method for processing non-circular signals, namely a parallel ACLMS adaptive filtering method. The method comprises the steps: respectively passing input signals through two ACLMS filters to respectively obtain outputs of the two filters; combining the outputs of the two filters through the optimal correlation coefficient sequenceto obtain a final output; updating the weight vectors of the two filters by using the outputs of the two filters; and updating and outputting repeated iteration to a steady state. According to the method, the limitation of a conventional adaptive filtering algorithm is overcome, the output estimation error can be reduced while the filtering convergence speed is increased, and the filtering efficiency and performance are greatly improved. When the method is used for processing non-circular signals, the processing effect is better than that of a common filter algorithm and a conventional filterparallel structure, the method has extremely high sensitivity and extremely high adaptability to the non-circular signals, and the method has excellent performances in convergence speed and convergence errors.

Description

technical field [0001] The invention belongs to the field of array signal processing, relates to an adaptive filtering algorithm technology, in particular to a parallel enhanced CLMS adaptive filtering method for processing non-circular signals. Background technique [0002] Adaptive filtering algorithm is an optimal filtering method developed in recent years. Because of its stronger adaptability and better filtering performance, it has been widely used in engineering practice, especially in information processing technology. In the adaptive filtering algorithm, there is often a contradiction between the convergence speed and the estimation error. Usually, a faster convergence speed means a larger estimation error, and a smaller estimation error means a slower convergence speed. In addition, the non-circular signal is a kind of signal with special nature of high-order data, and its high-order data will change with the rotation of the signal, which often appears in amplitude ...

AI Technical Summary

Problems solved by technology

[0003] Aiming at the problem that the adaptive filter cannot take into account the convergence speed and estimation error, and the performance is not good when dealing with non-circular signals, the present invention provides a parallel enhanced CLMS adaptive filtering method for processing non-circular signals, and designs a This filter parallel structure is composed of two ACLMS filters with different convergence step sizes, which makes the filtering process have both faster convergence speed and smaller estimation error; at the same time, the enhanced complex minimum mean square error filter is used in parallel In the structure, the obtained parallel filter structure has excellent performance for processing second-order non-circular signals

Images