| "System identification using the adaptive filter is commonly utilized in areas such as noise cancellation, telephone communication, control, radar system, etc... In this research, we are going to implement the Recursive Least Squares (RLS) algorithm with an adaptive filter to identify an unknown system on TMS320C6713 DSK (DSP Starter Kit) using Code Composer Studio v. 6.2. An unknown system is arranged in parallel with the adaptive filter and a common signal is applied to both filter and unknown system. The output from both systems is subject to summation and the error is fed back to adaptive filter where the RLS algorithm updates the filter coefficients and the process is repeated. When the error reaches its minimum value, the output of the adaptive filter represents the unknown system. Least Mean Square (LMS) algorithm is readily available for system identification; the issue with LMS is that the convergence rate of filter coefficients is low because the coefficients updated at a fixed rate. This issue is solved when we use RLS adaptive algorithm with a forgetting factor at the cost of greater computational complexity" ("System Identification using Adaptive Filter with Recursive Least Squares" [1]). |
