Nonlinear dynamic compensation of sensors using inverse-model-based neural network

Dongchuan Yu, Fang Liu, Pik Yin Lai, Aiguo Wu

研究成果: 雜誌貢獻期刊論文同行評審

18 引文 斯高帕斯(Scopus)


Many sensors (such as low-cost sensors), in essence, display strongly nonlinear dynamic behavior that cannot be calibrated by well-developed linear dynamic compensation methods. So far, no general nonlinear dynamic compensation (NLDC) method exists, although there are some approaches based on nonlinear models (including Volterra series expansion, Wiener kernels, the Hammerstein model, and finite impulse response) that were developed to compensate some special kinds of nonlinear sensors. In this paper, we suggest a general framework for NLDC, in which removal of the influence of disturbance by using an auxiliary sensor is significantly studied and presented. The inverse model and differential-estimation-filter arrays are embedded in this general framework, where a neural network is applied to approximate the inverse mapping, and differential-filter arrays are used to estimate signal differentials up to a certain order. We also discuss the existence conditions of the general framework. The detailed design procedure of this general method is given as well. Simulation and experiments are presented to illustrate the proposed general NLDC method.

頁(從 - 到)2364-2376
期刊IEEE Transactions on Instrumentation and Measurement
出版狀態已出版 - 2008


深入研究「Nonlinear dynamic compensation of sensors using inverse-model-based neural network」主題。共同形成了獨特的指紋。