This paper introduces an adaptive adjusting FIR filter’s parameters (LMS) method and presents a system recognition model based on the adaptive filter theory. The adaptive filter is directly with adjustable coefficient h (0),h (1),...h (N-1).The unknown system and FIR model have the same input sequence. The simulation result confirms the feasibility of the model.
Separating the Effects of Batch Normalization on CNN Training Speed and Stability Using Classical Adaptive Filter Theory