摘要
针对具有多传感器量测的动态系统状态有效估计的问题,给出了一种基于多传感器数据融合的鲁棒自适应滤波算法。引入了虚拟量测的概念,实现对于量测数据的预处理,依据模糊集合理论中隶属度的性质定义了一种无需门限设定的数据间置信度函数,以充分提取于测量数据中互补和冗余信息。最后结合卡尔曼滤波中的贝叶斯递推估计原理,完成对于量测方差的在线自适应估计,进而提高了系统状态的滤波精度,并通过计算机仿真验证了算法的有效性。
Aiming at the issue of efficient state estimation for dynamic system with multiple sensors, a robust self-adaptive filtering algorithm based on multi-sensor data fusion is given. In the algorithm, the concept of virtual measurement is introduced for pre-processing of the measured data. In accordance with the feature of subordinate degree in fuzzy set theory, a confidence function among data without limit setting has been designed. Thus complementary and redundant information in measured data are fully extracted. Combining with the Bayesian recursive estimation principle in Kalman filtering, the on-line self-adaptive estimation of the measurement variance is implemented. The computerized simulation verified the effectiveness of the algorithm.
出处
《自动化仪表》
CAS
2008年第6期65-67,75,共4页
Process Automation Instrumentation
基金
广东省森林土壤墒情自动监测和预报系统研究项目(编号:2007B030402001)
关键词
数据融合
虚拟量测
隶属度
置信度函数
自适应滤波
贝叶斯估计
Data fusion Virtual measurement Subordinate level Confidence function Self-adaptive filtering Bayes estimation
