Pulse or digital communications – Equalizers – Automatic
Reexamination Certificate
1999-12-09
2003-07-08
Bocure, Tesfaldet (Department: 2631)
Pulse or digital communications
Equalizers
Automatic
C708S301000
Reexamination Certificate
active
06590931
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention generally relates to digital filters, and more particularly to a finite-impulse response (FIR) filter that is reconfigurable for different sets of canonical sign digit (CSD) coefficients.
One technique for implementing FIR filters is the CSD representation. This technique provides a substantial advantage in the hardware implementation of FIR filters by eliminating the need for multipliers, which are the largest and most expensive components required to perform digital filtering operations. Instead, the CSD representation permits the minimization of the total number of non-zero bits in all of the filter coefficients that nevertheless maintains acceptable filter performance.
For each tap in a CSD representation filter, the data tap value is shifted by the number of bit positions corresponding to the position of each non-zero bit in the coefficient for that tap. The resulting shifted data tap values are then added. This is done for every coefficient in the filter. A primary consequence of this is that a larger adder (i.e., one with more data inputs) is required than would be needed for a conventional FIR filter implementation. However, this is still desirable since no multipliers are used. The high hardware cost of multipliers makes the CSD implementation especially attractive for very long digital filters. Such filters are often needed, for example, in communications systems, particularly for the demodulation of digital data signals.
In CSD filters, each coefficient is implemented by adding the associated data tap value shifted by a coefficient bit position. Since the shifting of the data tap values is usually a hard-wired function of the data bus, CSD filters are designed exclusively for a specific set of coefficients and thus cannot be reconfigured or programmed for different coefficients. This is in contrast with a conventional FIR filter, in which each coefficient is simply multiplied by the data tap value. Such a filter can easily be reconfigured if a programmable register is provided at the coefficient input of each data tap multiplier, so that every coefficient can be reconfigured to any desired value simply by programming that value into the register.
SUMMARY OF THE INVENTION
The present invention is directed to a reconfigurable or programmable FIR filter having coefficients that are represented using the canonical sign-digit (CSD) format. According to the present invention, the FIR filter can be reconfigured at any time by specifying (i.e., programming) different values for the coefficients, limited only by the constraints of maximum values for filter length, data and coefficient precision.
The FIR filter according to the present invention includes a shift group associated with at least one of the data tap values and at least one of the CSD coefficients. Each shift group includes a plurality of parallel paths. Each parallel path shifts the data tap value according to a bit pair of the coefficient to produce a weighted product. An adder is also included that combines the weighted products from each of the shift groups.
Further, each parallel path includes a shifting unit for shifting the data tap value a predetermined number of places corresponding to a position within the coefficient of a first bit of the bit pair. Each parallel path further includes an adder-input unit coupled to an output of the shift unit for further shifting the data tap value if a second bit of the bit pair is non-zero. The adder-input unit also passes the data tap value from the shifting unit unchanged to the adder if the second bit of the bit pair is zero. The adder-input unit also forces the data tap value from the shifting unit to zero if both bits of the bit pair are zero. The adder-input unit further inverts the data tap value from the shifting unit if one bit of the bit pair is negative.
The present invention is also directed to a method for processing a digital filter tap value with a canonical sign digit (CSD) coefficient. The method includes replicating the digital filter tap value to produce replicated tap values. The method also includes shifting each of the replicated tap values according to a bit pair of the CSD coefficient to produce weighted products and then adding the weighted products.
REFERENCES:
patent: 4967388 (1990-10-01), Tate
patent: 5081604 (1992-01-01), Tanaka
patent: 5262974 (1993-11-01), Hausman et al.
patent: 6282184 (2001-08-01), Lehman et al.
patent: 766388 (1997-04-01), None
LSI Logic Corp, Description, 1990 LSI Logic Corp, L64245 40 MHZ Versatile FIR Filter, pp. 1-12.
James M. Andersen, TRW, “Parameterizable CSD Multiplier for Digital Filter Applications”, SMP, View/Methodology Notes.
Chao-Liang Chen and A.N. Willson Jr., “Higher Order &Sgr;&Dgr; Modulation Encoding for Design of Multiplierless FIR Filters”, Electronics Letters, Nov. 26, 1998, vol. 34, No. 24, pp. 2298-2300.
M. Mehendale et al, “Coefficient Transformations for Area-Efficient Implementation of Multiplier-Less FIR Filters”, 1997 IEEE, 1063-9667/ 1997, pp. 110-115.
M. Yagyu et al, “Design of Linear Phase FIR Digital Filters Using Minimal Number of Adders and Subtractors”, 1996 IEEE, O-7803-3702-6/96, Proceedings of IEEE Asia Pacific Conf. on Circuits and Systems, 1996, Nov. 18-21, 1996, Seoul Korea, pp. 286-289.
W. Jin Oh and Y. Hoon Lee, “A Design and Implementation of Programmable Multiplierless FIR Filters with Powers of Two Coefficients”, Dept. of Electrical Eng. Korea Advanced Institute of Science and Tech., Proceedings 1993 IEEE Int'L Symposium on Circuits and Systems, pp. 88-91.
Bor-Rong Horng et al, “The Design of Low-Complexity Linear-Phase FIR Filler Banks Using Powers-Of-Two Coefficients with an application to Subband Image Coding”, IEEE Transactions on Circuits and Systems for Video Tech. vol. 1, No. 4, Dec. 1991, pp. 318-324.
Bocure Tesfaldet
Gross Russell
Koninklijke Philips Electronics , N.V.
Kumar Pankaj
LandOfFree
Reconfigurable FIR filter using CSD coefficient representation does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Reconfigurable FIR filter using CSD coefficient representation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Reconfigurable FIR filter using CSD coefficient representation will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3072596