人工神经网络在改善报警系统性能中的应用

摘要

介绍了一种利用BP人工神经网络降低报警系统误报警率的有效方法.以STC89C52RC处理器为平台,在嵌入式系统中应用BP网络先对报警信息进行分析和预处理,再控制报警系统动作.该方法大大降低了报警系统因自然因素导致的误报警概率.采集了100个典型数据对BP网络进行训练,得到了适合应用目标的网络模型,现场实测误报警率几乎为0.该方法的硬件价格便宜,电路简单,是一种降低报警系统误报警率的有效方法.%The author introduces an effective method which makes use of the BP network to reduce the mistakes of the alarm system. Based on STC89C52RC processor, the BP network in embedded systems analyses and pretreates the alarm information, then makes the alarm system act. This method greatly reduces the error rate because of natural factors. Using 100 typical samples data which have trained by BP network, the author obtained the suitable network model for application, the error rate is close to zero. This method has many advantages such as low price and simple circuit. It is an effective method to reduce the error rate the alarm systems.