New nonlinear adaptive FIR digital filter for broadband noise cancellation

The structure and the adaptive algorithm of a new nonlinear adaptive FIR digital filter. Called nonlinear delayed N-path adaptive FIR (NDNAFIR) digital filter, is presented for cancelling broadband noise from signals containing sharp edges. The filter structure is obtained by decomposing a FIR digital filter into N delayed parallel subfilters which facilitates high speed implementation. Adjustable bipolar sigmoid functions are placed at the output of each subfilter as well as after the summation of all subfilter outputs. Simulation results show that the ND-NAFIR digital filter outperforms those of the linear adaptive FIR digital filter and the nonlinear median filter in broadband noise cancellation in terms of convergence speed and average squared difference between original and recovered signals. >

[1]  Xiao-Hu Yu,et al.  Nonlinear statistical optimum adaptive filtering and signal detection via BP neural nets , 1991, 1991., IEEE International Sympoisum on Circuits and Systems.

[2]  Yannis Papananos,et al.  Analysis and VLSI architecture of a nonlinear edge-preserving noise-smoothing image filter , 1991 .

[3]  Bernard Widrow,et al.  Adaptive Signal Processing , 1985 .

[4]  Rolf Unbehauen,et al.  Edge preserving filtering by combining nonlinear mean and median filters , 1991, IEEE Trans. Signal Process..

[5]  S. Haykin,et al.  Adaptive Filter Theory , 1986 .

[6]  Akio Miyazaki,et al.  A new adaptation algorithm of adaptive digital filters using neural networks , 1991, 1991., IEEE International Sympoisum on Circuits and Systems.

[7]  Ioannis Pitas,et al.  Nonlinear Digital Filters - Principles and Applications , 1990, The Springer International Series in Engineering and Computer Science.