Identification of FIR Systems with Quantized Input and Binary-Valued Observations Under A Priori Parameter Constraint

This paper investigates the identification of FIR (finite impulse response) systems with quantized input and binary-valued observations. First, we obtain the ML (maximum likelihood) function of the available data, and construct an estimation algorithm of the unknown parameters when solving the maximum likelihood solution by transforming it into the solution to a set of linear equations. Secondly, based on the weighted least squares optimization technique, we establish the corresponding estimation algorithms in the case that the unknown parameters respectively satisfy a priori equality constraint and inequality constraint. Then, the AIC criterion is designed for estimating the order of the system. Finally, a numerical simulation example is employed to verify the effectiveness of the theoretical results obtained.