作者: Maciej Niedzwiecki , Szymon Gackowski
关键词: Estimation theory 、 Linear system 、 Algorithm 、 Mathematics 、 Basis function 、 Mathematical optimization 、 Technical note 、 Exponential growth 、 Smoothing 、 Finite impulse response 、 Bandwidth (signal processing)
摘要: In this technical note, we consider the problem of finite-interval parameter smoothing for a class nonstationary linear stochastic systems subject to both smooth and abrupt changes. The proposed parallel estimation scheme combines estimates yielded by several exponentially weighted basis function algorithms. resulting smoother automatically adjusts its bandwidth type rate nonstationarity identified system. It also allows one account distribution measurement noise.