Abstract: This paper aims the method of estimating unknown system characteristics (desired signal) by using the system identification method. The design of such system belongs to the optimal filtering domain, which is originated from the work of Wiener. In such applications fixed or adaptive Filters can be used. Fixed filters designed based on system prior knowledge, but adaptive algorithm based filters consists the ability to adjust their own parameters automatically and can match their (unknown system) characteristics with desired system (known system) characteristics. Design of these adaptive systems not required any prior knowledge of system characteristics. In the proposed system parameters were estimated using various least square adaptive algorithms. Since a finite impulse response(FIR) filter only has zeros, is stable irrespective of the filter coefficients, this work selects a FIR filter as the base filter. The adaptive algorithm used is Least Mean Square (LMS) and various LMS algorithms like Normalized-LMS, Signed-LMS, Signed–Signed LMS, Signed data, Signed error etc. The proposed work used C software to develop System identification with adaptive FIR filter using various LMS algorithms and also implemented this model by using TMS320C6713 DSK for Real Time Applications. Texas Instrument (TI) assembly language can be obtained by using 3.1V Code Composer Studio (CCS).

Keywords: Adaptive algorithms, estimation of unknown characteristics, LMS, NLMS, signed error LMS, signed-data LMS, sign-sign LMS algorithm.