摘要
针对常规卡尔曼滤波(KF)处理小噪声和变化噪声不足,提出了一种改进的自适应模糊卡尔曼滤波[1(]IAF-KF)算法。该算法根据模糊推理输入量的变化特点建立一个新的非线性隶属度函数,取代了常用的三角形线形隶属度函数;然后利用模糊化后的等级和隶属度构造了补偿调节函数(CAF),用于调节卡尔曼滤波算法中的误差,提高实际测量误差与理论测量误差间的匹配程度。仿真实验表明,较之传统的卡尔曼滤波,该方法在小噪声和变化的噪声情形下有效的克服了稳态误差,同时降低了模糊卡尔曼滤波算法的复杂程度。
An improved adaptive fuzzy kalman filter algorithm is presented aimed at the normal kalman filter deficiency to deal with minor noise and change noise.First a new nonlinear membership function is established based on the characteristic of the fuzzy inferential input,instead of normal triangular membership function.Then a compensated adjusted function is constructed with grade and membership to adjust kalman filter error.So the matching degree is improved between actual and theoretical measure error.Simulation result indicated that the algorithm is efficiency to control steady error and degraded the complexity of fuzzy kalman filter in minor noise and change noise.
出处
《计算机工程与应用》
CSCD
北大核心
2007年第28期25-28,共4页
Computer Engineering and Applications
基金
国家自然科学基金(the National Natural Science Foundation of China under Grant No.10672044)。
关键词
模糊推理
卡尔曼滤波
补偿调节函数
自适应调节
fuzzy inference
Kalman Filter
compensated adjusted function
adaptive adjust
