基于状态BPA折扣因子的过程状态估计算法  

Process State Estimation Algorithm Based on State BPA Discount Factor

作　　者：周理[1,2] 刘琰[1] 

机构地区：[1]福建工程学院信息科学与工程学院,福建福州350108 [2]福建省数字化装备重点实验室,福建福州350108

出　　处：《信息与控制》2014年第2期141-145,共5页Information and Control

基　　金：福建省高校产学合作科技重大项目(2010H6002;2013H6001)

摘　　要：为了提高过程控制系统中的状态估计精度,提出一种基于状态BPA(basic probability assignment)折扣因子的过程状态估计算法.首先,改进了状态BPA函数,引入元知识对证据信任度进行重新分配;然后为了设计重新分配的规则,定义了状态BPA折扣因子,该因子能够让证据距离达到最小;最后基于新提出的状态BPA折扣因子设计过程状态估计算法.对一个精密零件注射成型的过程监控进行仿真实验,结果表明:该算法可以有效消除测量不确定性对状态估计的影响,提高状态估计精度.To improve the state estimation precision of process control system,a process state estimation algorithm based on the state BPA（ basic probability assignment） discount factor is proposed.First,the measure function of state probability is improved.Then evidence confidence is redistributed by introducing a meta-knowledge.In order to design the reallocation rules,a state BPA discount factor is defined.The factor can make evidence distance reach the minimum.An estimation algorithm of process state is designed based on the new proposed state BPA discount factor.A process monitoring of injection molding is simulated.The results show that the algorithm can effectively eliminate measurement uncertainty on the influence of state estimation,and the precision of state estimation is improved.

关 键 词：过程状态 BPA(basic PROBABILITY assignment) 折扣因子 证据理论 

分 类 号：TP11[自动化与计算机技术—控制理论与控制工程]