摘要
介于振动筒压力传感器在航空领域有广泛的应用,本文以某基于振动筒压力传感器的大气数据系统为基础,研究了振动筒压力传感器的输入-输出特性,并对影响其输出特性的主要因素进行了深入分析,得到了振动筒压力传感器的温度特性曲线.提出了一例利用BP人工神经网络对振动筒压力传感器静态输出特性进行修正的新方法.计算机仿真和试验结果表明:该方法能够有效改善传感器的输出特性,并且速度快、精度高、鲁棒性强,便于用硬件实现,具有较高的推广应用价值.
As vibration-type sensors are widely used in navigation instruments, based on the theory of vibration-type sensor.The input/output of vibration-type sensors is deeply studied. Through analysis of the experimental data,the temperature characteristic curve is obtained. And a new method based on BP neural network’s non-linear function approach to improve the static state output characteristic of vibration-type sensor is presented. The simulation result demonstrates that the vibration-type sensor’s static state output characteristic has improved effectively. Besides that,simple hardware,faster speed,high accuracy and better robustness are offered.
出处
《传感技术学报》
CAS
CSCD
北大核心
2007年第10期2213-2217,共5页
Chinese Journal of Sensors and Actuators
基金
教育部科学技术研究重点项目资助(206077)
江西省教育厅基金项目资助(2006191)
关键词
BP神经网络
振简式传感器
大气数据系统
静态输出特性
鲁棒性
BP neural network
vibration-type sensor
air data system(ADS)
static state output characteristic
rebustness
