基于云模型构造BPA的瓦斯监测证据合成算法

摘要

D-S evidence combination can improve overall decision and early warning capability in coal mine gas monitoring system. In this study, the danger levels of local region in coal mine are defined according to coal mine safety specification. Cloud model is used for generating the curve clusters of the membership degrees corresponding to the danger levels. The membership degrees extracted from the curve clusters are regarded as basic probability assignment for D-S evidence combination. The study puts forward a step-by-step combination algorithm mixing D-S method and the weighted average method for severe conflicting evidence combination. The simulation results show that the proposed algorithm has satisfactory convergence effect on severe conflicting evidence combination.%煤矿瓦斯监测中，利用D-S证据合成方法实现多传感器信息融合可以提高系统整体决策和预警能力。根据煤矿安全规范设定区域危险等级，使用云模型建立危险等级属性隶属度曲线簇，输入传感器检测量提取各属性隶属度作为D-S融合的基本概率赋值。为了实现高度冲突证据合成，提出D-S与加权平均法混合的分步证据合成算法。仿真结果表明文中提出的算法合成高度冲突证据时，具有令人满意的收敛效果。