Mixed norm regularized recursive total least squares for group sparse system identification.

A mixed l p,0-regularized recursive total least squares (RTLS) algorithm is considered for group sparse system identification. Regularized recursive least squares (RLS) has been successfully applied to group sparse system identification; however, the estimation performance in regularized RLS-based algorithms deteriorates when both input and output are contaminated by noise (the error-in-variables problem). We propose an l p,0-RTLS algorithm to handle group sparse system identification with errors-in-variables. The proposed algorithm is an RLS-like solution that utilizes l p,0-regularization. The proposed algorithm provides excellent performance as well as reduces the required complexity by effective inversion matrix handling. Simulations demonstrate the superiority of the proposed l p,0-regularized RTLS for a group sparse system identification setting. Copyright © 2015 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]