基于距离矢量的多传感器信息决策算法  

Multi-sensor information decision algorithm based on distance vector

作　　者：张凌飞 马俊[1] 陈博行[1] ZHANG Lingfei;MA Jun;CHEN Bohang(School of Computer,Qinghai Normal University,Xining 810008,China;College of Physics and Electronic Information Engineering,Qinghai Nationalities University,Xining 810007,China)

机构地区：[1]青海师范大学计算机学院,青海西宁810008 [2]青海民族大学物理与电子信息工程学院,青海西宁810007

出　　处：《现代电子技术》2022年第1期30-33,共4页Modern Electronics Technique

基　　金：国家自然科学基金项目(61761040);教育部春晖计划(Z2015062);青海省重大科技专项(2019-ZJ-A10);青海省135高层次人才资助和青海省物联网重点实验室建设专项(2017-ZJ-Y21)。

摘　　要：当前大多传感器信息决策过程中均需要对传感器数据进行概率转换或进行权重计算,该过程不仅计算量大,且决策结果受概率值影响较大,故提出一种基于距离矢量的多传感器信息决策算法。该算法利用传感器信息决策特点,结合系统评价标准,通过对支持度矩阵的归一化处理,得出优势度函数,对优势度函数规范化后进行排序,取最大值为最优解方案。定义了区分度函数,为算法评价提供了科学依据。经仿真案例分析表明,该方法在传感器信息决策中识别区分度均达到了0.5以上,决策效果良好。通过与经典D-S证据理论进行比较,可有效避免D-S证据理论在决策时出现与实际相悖的情况,且相对于D-S证据理论在决策过程中需对初始传感器数据进行概率赋值,具有明显优势。At present,most of the sensor information decision-making processes need to perform probability conversion or weight calculation for the sensor data. As far as the decision-making process is concerned,its computation burden is heavy,and its decision result is greatly influenced by the probability value. Therefore,a multi-sensor information decision algorithm based on distance vector is proposed. In this algorithm,in combination with the system evaluation standards,the characteristics of sensor information decision-making is adopted to obtain the dominance function by normalizing the support matrix. The dominance functions are sorted after normalization,and its maximum value is the optimal solution. The discrimination function is defined,which provides a scientific basis for algorithm evaluation. Simulation case analysis shows that the method has a recognition discrimination of more than 0.5 in sensor information decision-making,with good decision-making effect. In comparison with the classical D-S evidence theory,the proposed method can effectively avoid the situation that D-S evidence theory contradicts the reality in decision-making. In the D-S evidence theory,the initial sensor data needs to be assigned a probability value in the decision-making process,so the proposed method has an obvious advantage.

关 键 词：信息决策 评价标准 概率赋值 距离矢量 支持度矩阵 归一化处理 优势度函数 区分度函数 

分 类 号：TN919-34[电子电信—通信与信息系统]