基于BP神经网络和D-S证据理论的火灾报警系统

Fire Alarm System Based on BP Neural Network and D-S Evidence Theory

将BP神经网络与数据融合理论中的D-S证据有机融合,提出一种决策级火灾报警识别方法。通过模拟实际输入信号的仿真结果发现,将BP神经网络和D-S证据理论相结合的多传感器数据融合技术,可以显著提高火灾的识别能力,有效降低火灾误报率,而且该系统具有良好的适应性,达到了预期效果。

This paper proposed a decision-level data fusion fire alarm identification method,which combined BP neural network with D-S theory of evidence.Through the simulation result of input signal,the technology based on BP neural network with D-S theory of evidence can improve the fire discrimination capability,and reduce the false positive rate of fire effectively.The system has good adaptive capacity,and reaches the expected results.