ANALYSIS OF EEG SIGNALS WITH CHANGING SPECTRA USING A SHORT-WORD KALMAN ESTIMATOR
ANALYSIS OF EEG SIGNALS WITH CHANGING SPECTRA USING A SHORT-WORD KALMAN ESTIMATOR
复制标题
DOI：
10.1016/0025-5564(77)90026-8
复制
发表时间：
1977-01-01
影响因子：
4.3
通讯作者：
BOHLIN, T
中科院分区：
文献类型：
--
作者：
BOHLIN, T
本文提出了一种新的脑电信号数值分析方法，该方法具有谱分布随时间变化的特点。变化不需要很慢。该理论用随机系数独立增量的高阶自回归序列描述采样脑电信号。这些系数作为状态变量被增广，因此可以通过一个特殊的卡尔曼滤波器进行真实的实时估计。这提出了两个问题。黎卡提方程依赖于信号，在数值上是不稳定的。调谐滤波器的规格是未知的;基本参数是随机系数的变化率。对于解决方案，一个数值稳定的算法推导（不是平方根滤波），并根据最大似然估计的变化率。这定义了一个指数的非平稳性，其价值是基本的分析。记录的EEG信号[人类]的应用程序证明了该方法的可行性。
A new method was described for the numerical analysis of EEG signals, whose spectral distributions vary with time. The variation need not be slow. The theory described a sampled EEG by an autoregressive series of high order with stochastic coefficients having independent increments. The coefficients were augmented as state variables and could thus be estimated in real time by a special Kalman filter. This raised 2 problems. The Riccati equation depended on the signal and was numerically unstable. Specifications for tuning the filter were unknown; the essential parameter was the rate of change of the stochastic coefficients. For solutions, a numerically stable algorithm was derived (not square-root filtering), and the rate of change was estimated according to maximum likelihood. This defined an index of nonstationarity, the value of which is fundamental for the analysis. Applications to recorded EEG signals [human] demonstrated the feasibility of the method.