Abstract
Usually, we deal with noise only in output side of a system for a system identification. However, it is known that there are algorithms using an input and output side noise for more appropriate estimation. Then, this paper presents two algorithms of a blind identification for a system in which there are both noises in input and output. The first is based on a simple adaptive gradient method. The second is based on RLS algorithm. When an output noise is added into system, we can achieve more precise estimation of system parameters. Computer simulations show our algorithms' performance.
