| Withthe continuous development of semiconductor technology, silicon-based piezoresistive pressure sensor has been widely used in various fields. Piezoresistive pressure sensors typically use in microelectronics pressure conduction film sputtered piezoresistive effect of semiconductor material having a composition of resistance varies with the pressure to achieve pressure measurement bridge, so there is often a pressure sensor measuring range nonlinearity and temperature drift and other issues. The practical application of silicon piezoresistive pressure sensors required for related compensation. Therefore, the nonlinear problems of Silicon piezoresistive pressure sensors for practical use has important significance.In order to improve the non-linearity of pressure sensor and the output error characteristics caused by the variation of temperatures,the piezoresistive pressure sensor’s circuit model structural model and the reason of nonlinear has given and tree methods for polynomial curve fitting,BP neural network arithmetic and GRNN neural network arithmetic has discussed.Aiming at abundent IO resource tight in large-scale compensating system,FPGA used for large-scale compensating system scheme is proposed and 32-bit single-precision FPU for realizing the polynomial curve fitting methods is designed.which lays the foundation for FPGA compensating system.The results show that the polynomial curve fitting methods that FPGA can realizeand Matlab can realize is consistent;GRNN neural network that based on RBF(Radial Basis Function) has many advantages such as fast convergence speed,simple structure and applicable microcontroller programming and so on.Overcome the BP neural network training complex, it is easy to fall into local optimum shortcomings.Temperature compensation from the final results, GRNN compensation algorithm effectively inhibit the effect of temperature on silicon piezoresistive pressure sensor output, improve the stability and accuracy of piezoresistive pressure sensors. |
PDF Full Text Request