Research And Implementation On Calibration And Compensation Of A Sensor System Based On RBF Network

Sensing technology is the foundation of information science. Sensor technology is one of the important footstones of the modern information technology, which has almost permeated through every corner of science and technology and national economy. Sensor network is a measurement and control system linked by mounts of sensor nodes communicated through certain wired or wireless protocols. In order to improve the precision of the sensors and the intelligent intensity in the wireless sensor network system, this paper applies RBF neural network to the calibration and temperature compensation of piezoresistance sensor to implement two functions. The first one is to obtain the nonlinear characteristics. The second one, on the other hand, is to calibrate the nonlinear characteristics of the piezoresistance sensor for the purpose of greatly reducing the sensor and wireless sensor network on the influence of the ambient temperature and the power fluctuation.On the basis of studying the determined method of RBF center, first, this paper concentrates on analyzing the dynamic nearest neighbor algorithm, summarizes its deficiency, and provides an improved dynamic adaptive nearest neighbor algorithm. Based on such algorithm, IDARBF neural network is constructed. Second, its generalization ability through the concrete function approximation example is analyzed and a method for obtaining better generalization ability through dynamic Gauss width searching is presented. The simulated results indicate that the improved nearest neighbor algorithm has better approaching ability and can obtain better network performance. Third, this IDARBF neural network is applied to piezoresistance sensor. The calibration and temperature compensation function are implemented through two neural networks modeling. The direct modeling technology enables an estimate of the nonlinear sensor characteristics, whereas the inverse modeling technology estimates the applied pressure. Fourth, after comparing the pressure fusion performance with the original adapted neural network and SLFM neural network, the simulated results show the IDARBF neural...