Adaptive Filters - Digital Signal Processing and Applications with the C6713 and C6416 DSK - page 270

( b )

FIGURE 7.18. ( Continued )

Example 7.4: Adaptive FIR Filter for System ID of a Fixed FIR as an
Unknown System ( adaptIDFIR )

Figure 7.19 shows a listing of the program adaptIDFIR.c , which models or iden-
tifies an unknown system. See also Examples 7.2 and 7.3, which implement an adap-
tive FIR for noise cancellation.
To test the adaptive scheme, the unknown system to be identified is chosen as an
FIR bandpass filter with 55 coefficients centered at F s /4 2 kHz. The coefficients

of this fixed FIR filter are in the file bp55.cof , introduced in Chapter 4. A 60-
coefficient adaptive FIR filter models the fixed unknown FIR bandpass filter.
A pseudorandom noise sequence is generated within the program (see Exam-
ples 2.16 and 4.4) and becomes the input to both the fixed (unknown) and the adap-
tive FIR filters. This input signal represents a training signal. The adaptation process
continues until the error signal is minimized. This feedback error signal is the dif-
ference between the output of the fixed unknown FIR filter and the output of the
adaptive FIR filter.
An extra memory location is used in each of the two delay sample buffers (fixed
and adaptive FIR). This is used to update the delay samples (see method B in
Example 4.8).
Build and run this project as adaptIDFIR . Verify that the output ( adapt-
fir_out ) of the adaptive FIR filter converges to a bandpass filter centered at 2 kHz
(with the slider in position 1 by default). With the slider in position 2, verify the

=