作者: S.A. Billings , M. El Hitmy , A.R.D. Curtis
DOI: 10.1016/0022-460X(90)90761-N
关键词:
摘要: Abstract A new filter structure known as the NARMAX model (Non-linear Auto Regressive Moving Average with eXogenous inputs) is introduced a basis for design of class non-linear filters which can be used noise cancellation. The an extension classical IIR filter, and it shown that parameters estimated by using suboptimal least squares algorithm originally developed system identification. Conditions convergence are briefly discussed simulation results presented show effectiveness structure.