Abstract:Separable nonlinear models (SNMs) serve as potent instruments for system identification, data analysis, and machine learning. However, the online identification of SNMs poses a greater challenge compared to their offline counterparts, primarily due to the dynamic nature of nonlinear systems. Traditional approaches, including the recursive Gauss–Newton and recursive variable projection methods, falter in managing parameter interdependence, culminating in sluggish convergence and suboptimal outcomes. Addressing these limitations, our study introduces a pioneering multi-innovation-based recursive variabl projection (MIRVP) algorithm, which extends the RVP algorithm with a multi-innovation strategy. This strategy enables the algorithm to handle the interaction of the linear and nonlinear parameters more effectively, by using multiple past innovations instead of only the current one, thus improving the identification effect. The proposed algorithm's effectiveness has been validated through tests on synthetic data, real-life industrial control tracking scenarios such as the Box–Jenkins gas furnace data and Glass tube drawing processing, as well as on the training of large-scale neural networks. The results demonstrate that our algorithm outperforms existing methods in terms of convergence speed and robustness

Auxiliary model based multi-innovation algorithms for multivariable nonlinear systems

Parameter identification of a class of nonlinear systems based on the multi-innovation identification theory

Multi-innovation Generalized Extended Stochastic Gradient Algorithm for Multi-Input Multi-Output Nonlinear Box-Jenkins Systems Based on the Auxiliary Model

The Innovation Algorithms for Multivariable State‐space Models

Hierarchical Least-Squares Based Iterative Identification for Multivariable Systems with Moving Average Noises

Multivariable CAR-like System Identification with Multi-innovation Gradient and Least Squares Algorithms

Multi-innovation Stochastic Gradient Algorithms for Input Nonlinear Time-Varying Systems Based on the Line Search Strategy

Gradient-based parameter estimation for input nonlinear systems with ARMA noises based on the auxiliary model

Auxiliary model maximum likelihood gradient‐based iterative identification for feedback nonlinear systems

Multi-innovation identification method for time-varying systems

Additional fractional gradient descent identification algorithm based on multi-innovation principle for autoregressive exogenous models

Multi-innovation stochastic gradient identification method

Two‐stage auxiliary model gradient‐based iterative algorithm for the input nonlinear controlled autoregressive system with variable‐gain nonlinearity

Multi‐innovation gradient‐based iterative identification methods for feedback nonlinear systems by using the decomposition technique

Auxiliary model‐based iterative parameter estimation for a nonlinear output‐error system with saturation and dead‐zone nonlinearity

Multi-innovation-based online variable projection algorithm for a class of nonlinear models

An Auxiliary Model Based Multi-Innovation Recursive Least Squares Estimation Algorithms for MIMO Hammerstein System

Auxiliary Variable-Based Identification Algorithms for Uncertain-Input Models

Maximum likelihood based multi‐innovation stochastic gradient identification algorithms for bilinear stochastic systems with ARMA noise

Performance analysis of coupled multi-innovation stochastic gradient identification method

Multi-innovation Parameter Identification Based on Coupling Auxiliary Model for Three-axis Turntable