Auxiliary Model-based Continuous Mixed p-norm Algorithm for Output-error Moving Average Systems Using the Multi-innovation Optimization

This paper discusses the parameter estimation problems for the output-error moving average (OEMA) systems under stochastic environments. The estimation problems with unknown inner variables and unmeasurable noise terms existed in the information vector are solved by the auxiliary model framework. Meanwhile, the algorithms utilize the continues mixed p-norm (CMPN) method to control the proportions of the error norms, which take into account each p-norm of errors for 1 ⩽ p ⩽ 2. To improve the identification accuracy further, a multiinnovation CMPN optimization algorithm is developed by expanding the scalar innovation to the innovation vector. The proposed optimal algorithms offer faster tracking speed and can obtain higher parameter estimation accuracy for both stochastic white noise and α-stable noise. Two examples of identification of OEMA systems are given to validate the theoretical analysis.