IJCAS

This paper studies the wave peak frequency identification using the stochastic gradient algorithm. The identification model is derived by using the wave disturbance model, and an extended stochastic gradient algorithm is presented for identifying the model parameters. Based on the hierarchical identification principle, the identification model is decomposed into two subsystems by introducing two intermediate variables, and a two-stage auxiliary model based extended stochastic gradient (2S-AM-ESG) algorithm is presented to improve the convergence speed. The effectiveness of the identification algorithms is verified by the simulation tests of a ship heading control system. Simulation results demonstrate the effectiveness of the proposed method.

Keywords: Auxiliary model identification, hierarchical identification, stochastic gradient, wave frequency tracker.