Two-stage adaptive compensation of active noise interference with signals orthogonalization of a part of compensation channels

Abstract

Introduction. Method of least squares (MLS) is widely spread for synthesis of digital automatic compensators of active noise interference. This method allows solving the problem of adaptive spatial noise filtration under conditions of expected uncertainty. However, the digital automatic compensators obtained with the help of LS- and RLS-algorithms do not allow implementation of parallel processing and require considerable computational cost. Also, they do not provide the possibility to connect/disconnect additional compensation channels or to change the structure of processing units adequately to the interference situation.
Problem statement. Autoregressive model is used for description of the interfering signals at the input of the digital multi-element antenna array. In this model regressors’ matrix is divided into two blocks. Columns of regression matrix (regressors) are iteration readings of complex enveloping interfering signals delivered for input to compensation channels. It is required to synthesize optimal by the least squares criterion digital automatic compensator, in this digital automatic compensator first interference compensation is performed in the main channel using one block of regressors, then the residue interference is compensated using the second block of regressors.
Synthesis of two-stage digital automatic compensators using LS-algorithm. Two-stage digital automatic compensators of active interference noise were synthesized using theorem of regression partition into two blocks (FWL-theorem). These compensators during the first stage compensate interference in the main channel with the use of the first block of regressors and orthogonalize regressors of the second block. During the second stage the compensators compensate residues interference with the orthogonalized regressors. The computing modules of the first and second stages perform orthogonal projection operations based on the block LS-algorithm.
Synthesis of two-stage digital automatic compensators using RLS-algorithm. Procedure of orthogonal projection performed in units of the first and second stages is represented in an equivalent form using weight coefficients. Recursive RLS-Kalman type algorithm is applied for their definition. In the optimal two-stage automatic compensators based on LS- and RLS- algorithms the processing of complete package of the input data is carried out primarily in the first stage and only then in the second stage.
Effectiveness analysis of the two-stage recurrent digital automatic compensator. An analysis of the effectiveness of the two-stage recurrent digital compensator based on RLS-algorithm with simultaneous operation of the stages is performed with the help of statistical modeling at different conditioning of the correlational interference matrix and different numbers of interference sources. When the second stage starts with a delay of a few tick marks the resulting compensator provides performance indicators close to the potentially achievable.
Conclusions. The synthesized two-stage active noise interference digital automatic compensators based on LS- and RLS-algorithms allow implementation of parallel processing. The resulting two-step recursive digital automatic compensator with simultaneous operation of the stages provides the possibility to connect / disconnect the compensation channels block and the relevant stage computing modules adequately to the interference situation.