Canonical pipelining of lattice filters

A variant of the lattice form is proposed in which delay elements are placed on both the upper (feedforward) path and the lower (feedback) path alternately. Although this new form preserves the fundamental characteristics of the usual lattice form having all the delay elements on the lower (feedback) path, it has some practical advantages. First, it represents a completely pipelined systolic-type array processor. The pipelining is canonical in the sense that no extra delay is introduced, Second, the finite word length performance is shown to be superior to that of the usual lattice form when it is used to implement digital filters of general transfer functions. The theoretical reasoning for this improvement is discussed based on the state-space representation. It is also shown that the new lattice form is topologically equivalent to the wave digital filter composed of the unit elements, which gives the physical ground of this new form.

[1]  H. Kimura,et al.  Generalized Schwarz form and lattice - ladder realizations of digital filters , 1985 .

[2]  A. Gray,et al.  Roundoff noise characteristics of a class of orthogonal polynomial structures , 1975 .

[3]  B. Liu,et al.  Error analysis of digital filters realized with floating-point arithmetic , 1969 .

[4]  A. Gray,et al.  A normalized digital filter structure , 1975 .

[5]  Clifford T. Mullis,et al.  Synthesis of minimum roundoff noise fixed point digital filters , 1976 .

[6]  Alfred Fettweis,et al.  Digital circuits and systems , 1984 .

[7]  Thomas Kailath,et al.  A view of three decades of linear filtering theory , 1974, IEEE Trans. Inf. Theory.

[8]  M. Morf,et al.  State-space structures of ladder canonical forms , 1980, 1980 19th IEEE Conference on Decision and Control including the Symposium on Adaptive Processes.

[9]  Lajos Gazsi,et al.  Explicit formulas for lattice wave digital filters , 1985 .

[10]  A.V. Oppenheim,et al.  Analysis of linear digital networks , 1975, Proceedings of the IEEE.

[11]  P. P. Vaidyanathan,et al.  A unified approach to orthogonal digital filters and wave digital filters, based on LBR two-pair extraction , 1985 .

[12]  Patrick Dewilde,et al.  Schur recursions, error formulas, and convergence of rational estimators for stationary stochastic sequences , 1981, IEEE Trans. Inf. Theory.

[13]  J. Makhoul.  A class of all-zero lattice digital filters: Properties and applications , 1978 .

[14]  Sun-Yuan Kung,et al.  On supercomputing with systolic/wavefront array processors , 1984 .

[15]  B. Friedlander.  Lattice methods for spectral estimation , 1982, Proceedings of the IEEE.

[16]  T. Kailath,et al.  On a generalized Szegö- Levinson realization algorithm for optimal linear predictors based on a network synthesis approach , 1978 .