Research on dynamic characteristics of multi-sensor system in the case of cross-sensitivity

In this study, the dynamic characteristic of a multi-sensor system made up of such sensors as are sensitive to several parameters is discussed, and the effect of cross-sensitivity on the precision of a measurement system is also discussed. A multi-sensor system is looked as a serial of a linear filter and a memoryless nonlinear system, i.e. Wiener system, and the subsequent information fusion system is regarded as a Hammerstein system, i.e. a serial of a memoryless nonlinear system and a linear filter. On the basis of static calibration, it is presented to determine the inverse filter in a Hammerstein system using blind deconvolution. In order to control the uncertainty of amplitude of signals recovered by blind deconvolution well, a regulation approach to regulating the inverse linear filter coefficient matrixes is presented according to the identity between inverse filter coefficient matrixes and static calibrating matrix. So the approximate inverse dynamic model of multi-sensor system is obtained, the degree of distortion of dynamic measurement result is reduced, the measurement precision is improved, and the need of practice can be reached. Simulation example and simulation result show that the recovered error of the inputs of sensor system, the frequency of which is 1/10 of sampling frequency, is 1/20 of the measurement results without dynamic compensation, and is one half of the measurement results with sole dynamic compensation, and the rapidity is improved 2 times. The dynamic compensation results of a metal oxide semiconductor methane sensor show that the dynamic measurement error is less than one half of that without dynamic compensation. So this method expands the bandwidth of multi-sensor system.

[1]  Yong Zhang,et al.  Cross sensitivity reduction of gas sensors using genetic algorithm neural network , 2001, SPIE Optics East.

[2]  Andrzej Cichocki,et al.  Stability Analysis of Learning Algorithms for Blind Source Separation , 1997, Neural Networks.

[3]  Ah Chung Tsoi,et al.  Blind deconvolution of dynamical systems using a balanced parameterized state space approach , 2003, 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, 2003. Proceedings. (ICASSP '03)..

[4]  Enzo Baccarelli,et al.  A new approach based on "soft statistics" to the nonlinear blind-deconvolution of unknown data channels , 2001, IEEE Trans. Signal Process..

[5]  Pierre Comon,et al.  Independent component analysis, A new concept? , 1994, Signal Process..

[6]  Kenji Fukumizu,et al.  Adaptive natural gradient learning algorithms for various stochastic models , 2000, Neural Networks.

[7]  Aapo Hyvärinen,et al.  Nonlinear independent component analysis: Existence and uniqueness results , 1999, Neural Networks.

[8]  Chong-Yung Chi,et al.  Cumulant-based inverse filter criteria for MIMO blind deconvolution: properties, algorithms, and application to DS/CDMA systems in multipath , 2001, IEEE Trans. Signal Process..

[9]  Junhua Liu,et al.  Cross sensitivity reduction of gas sensors using genetic neural network , 2002 .

[10]  Shun-ichi Amari,et al.  Novel On-Line Adaptive Learning Algorithms for Blind Deconvolution Using the Natural Gradient Approach , 1997 .

[11]  Alisa Rudnitskaya,et al.  Cross-sensitivity of chalcogenide glass sensors in solutions of heavy metal ions , 1996 .

[12]  Doron Lancet,et al.  A feature extraction method for chemical sensors in electronic noses , 2003 .

[13]  Ho Chang,et al.  Analysis of dynamic characteristics of pressure sensors using square pressure wave theory and system identification , 2003 .

[14]  Tian She.  Application of Recurrent Network Model on Dynamic Compensation of Sensors , 2003 .

[15]  W. Minkina,et al.  Fast temperature determination using two thermometers with different dynamical properties , 2002 .

[16]  Sergio Cruces,et al.  On a new blind signal extraction algorithm: different criteria and stability analysis , 2002, IEEE Signal Processing Letters.