New Recursive Estimation Algorithms for Nonlinear System

The foundation of both system analysis and controller design is mathematica modeling. This paper studies the parameter estimation problem of the nonlinear system. An investigation is conducted into a hierarchical stochastic gradient approach by utilizing the hierarchical identification. The proposed method improves the accuracy of parameter estimation by using the multi-innovation identification. Increasing the innovation length is essential to improve the speed of convergence. A simulation demonstrates the effectiveness of the method.