神经网络瓦斯传感器及其在系统中应用.pdfVIP

上海大学 2010～2011 学年冬季学期研究生课程 课程名称： 新型传感技术 课程编号： 091102908 论文题目: 神经网络瓦斯传感器及其在系统中的应用 研究生姓名: 刘金鼎 学 号: 论文评语: 成 绩: 任课教师: 顾惠芬 评阅日期: 神经网络瓦斯传感器及其在系统中的应用 刘金鼎 （上海大学机电工程与自动化学院，上海 200072） 摘要：瓦斯传感器的工作环境十分恶劣，传感器的输出不仅仅取决于目标参量——瓦斯，还与非目标参量（温度、 湿度）等有关。鉴于神经网络良好的非线性逼近能力，将神经网络引入瓦斯传感器，建立智能传感器系统可以一 定程度上改善瓦斯传感器的非线性。本文采用多个传感器同时监测目标参量和其它非目标参量，其输出数据经 BP 神经网络和 RBF 神经网络融合处理，从而消除环境因素变化的影响，提高目标传感器的性能。实验仿真结果表明, 该方法行之有效,为煤矿安全监测提供了一种新的思路。 关键词：瓦斯传感器；神经网络；非线性 Neural Network Gas Sensor and its Application in System LIU Jin-ding (School of Mechatronics Engineering Automation, Shanghai University, Shanghai 200072, China) Abstract: The working environment of gas sensor is very poor, so the output of the sensor is not only decided by the target parameter, but also affected by other non-target parameters, such as temperature, humidity and so on. Because the nerve network has good non-linearity approaching ability, introducing nerve network into gas sensor and establishing the intelligent sensor system can improve the nonlinearity of gas sensor in the certain degree. The study applied the technology which the target parameter and the non-target Parameters are monitored by multi-sensors, then the data for the output is send to confusion through BP and RBF neural network technology, which can eliminate the effect of the variation for the environmental factors and improve the function of the sensor. Simulation results show that the effective method provides a new way of thinking for coal mine safety monitoring. Key words: gas sensor; artificial neural network; nonlinearity 1 前言 1.1 智能传感器的发展 当今时代是信息化的时代，各个领域常以信息的获取与利用