Research On Array-Type Wireless Pressure Sensor System Based On Neural Network Compensation

MEMS silicon piezoresistive pressure sensor has the advantages of low cost,small size and easy to be manufactured, as a result of which it has been widely used in meteorology, industrial production, petrochemical industry, biological medicine,and so on. However, the traditional piezoresistive pressure sensor has low measurement accuracy, which is mainly attributed to the change of the temperature of the environment could lead to the random errors including significant thermal drift of the zero point output and the sensitivity, as well as the poor stability, the obvious hysteresis nonlinearity and the repeatability. In recent years, with the rapid development of pressure sensor technology, the market has a higher requirements for it.In order to meet the needs of the market, a silicon piezoresistive pressure measurement system with good accuracy and stability should be designed. In this work, the hardware and software compensation algorithm of the traditional pressure measurement system based on MEMS silicon piezoresistive pressure sensor is improved. Firstly,the stability of the pressure measurement system is improved by using the GE NovaSensor NPC-1210-015A-3L MEMS pressure sensors within a thousandth of accuracy to constitute a 2x4 sensor array and the ensor array to reduce the random error caused by the creep of the sensor. Secondly, the output voltage signal's acquisition processing of the array pressure sensor system and the temperature sensor for compensation is realized by the AD7794 conversion, while STM32F103 is slected as the main control chip. Besides, in the software compensation of pressure sensor, the wavelet neural network algorithm based on quantum behaved particle swarm optimization is applied in temperature compensation for output signal, and then using the least squares method and T-S fuzzy neural network model for hysteresis and repeatability of the further error compensation, which improves the measurement accuracy of the system. Finally, in order to facilitate observing and recording data, the APP based on Andriod platform is developed. The compensated and revised data are sent to the mobile phone receiver by the wireless Bluetooth transmission module, and mobile phone platform as a host computer achieves the reception and display of wireless data, statistics, as well as monitoring and measuring for user.The temperature compensation range of the arrayed wireless high precision silicon piezoresistive pressure sensor system is 20?50? . Experimental calibration and test analysis found that the overall error of the system was reduced from± 0.152% to ± 0.058% .The results show that the temperature drift error,hysteresis nonlinearity and repeatability error of the piezoresistive sensor are effectively suppressed by hardware optimization and neural network algorithms, which can meet the higher market requirements.