A Novel DCGA Optimization Technique for Guaranteed BIBO-Stable Frequency-Response Masking Digital Filters Incorporating Bilinear Lossless Discrete-integrator IIR Interpolation Sub-Filters

This work is concerned with the development of a novel diversity-controlled (DC) genetic algorithm (GA) for the design and rapid optimization of frequency-response masking (FRM) digital filters incorporating bilinear lossless discrete-integrator (LDI) IIR interpolation sub-filters. The selection of FRM approach is inspired by the fact it lends itself to the design of practical sharp-transition band digital filters in terms of gradual-transition band FIR interpolation sub-filters. The proposed DCGA optimization is carried out over the canonical-signed-digit (CSD) multiplier coefficient space, resulting in FRM digital filters which are capable of direct implementation in digital hardware. A novel CSD look-up table (LUT) scheme is developed so that in every stage of DCGA optimization, the IIR interpolation sub-filters constituent in the intermediate and final FRM digital filters are guaranteed to be automatically BIBO stable. The proposed DCGA optimization permits simultaneous optimization of the magnitude-frequency as well of the group-delay frequency response of the desired FRM digital filters. An example is given to illustrate the application of the resulting DCGA optimization to the design of a lowpass FRM digital filter incorporating a fifth-order bilinear-LDI interpolation subfilter.

EFFICIENT DESIGN OF NARROWBAND COSINE-MODULATED FILTER BANKS USING A TWO-STAGE FREQUENCY-RESPONSE MASKING APPROACH

A new cosine-modulated filter bank (CMFB) structure is proposed based on the frequency-response masking (FRM) approach using masking filter decomposition. The resulting structure, the so-called FRM2-CMFB, presents reasonable computational complexity (number of arithmetic operations per output sample) and allows one to design filter banks with extremely large number of bands. The examples include the use of M=1024 bands, where the standard minimax method cannot be employed. These examples indicate that the reduction in computational complexity can be as high as 60% of the original FRM-CMFB structure, which does not use masking filter decomposition.

A MODIFIED FREQUENCY-RESPONSE MASKING STRUCTURE FOR HIGH-SPEED FPGA IMPLEMENTATION OF SHARP FIR FILTERS

This paper presents the design and implementation of high-speed, multiplierless, arbitrary bandwidth sharp FIR filters based on frequency-response masking (FRM) technique. The FRM filter structure has been modified to improve the throughput rate by replacing long band-edge shaping filter in the original FRM approach with two to three cascaded short filters. The proposed structure is suitable for FPGA as well as VLSI implementation for sharp digital FIR filters. It is shown by an example that a near 200-tap equivalent Remez FIR filter can be implemented in a single Xilinx XC4044XLA device that operates at sampling frequency of 5.5 MHz.

OPTIMIZATION OF FREQUENCY-RESPONSE MASKING BASED FIR FILTERS

A very efficient technique to drastically reduce the number of multipliers and adders in implementing linear-phase finite-impulse response (FIR) digital filters in applications demanding a narrow transition band is to use the frequency-response masking (FRM) approach originally introduced by Lim. The arithmetic complexity can be even further reduced using a common filter part for constructing the masking filters originally proposed by Lim and Lian. A drawback in the above-mentioned original FRM synthesis techniques is that the subfilters in the overall implementations are separately designed. In order to further reduce the arithmetic complexity in these two FRM approaches, the following two-step optimization technique is proposed for simultaneously optimizing the subfilters. At the first step, a good suboptimal solution is found by using a simple iterative algorithm. At the second step, this solution is then used as a start-up solution for further optimization being carried out by using an efficient unconstrained nonlinear optimization algorithm. An example taken from the literature illustrates that both the number of multipliers and the number of adders for the resulting optimized filter are less than 80% compared with those of the FRM filter obtained using the original FRM design schemes in the case where the masking filters are separately implemented. If a common filter part is used for realizing the masking filters, then an additional reduction of more than 10% is achieved compared with the optimized design with separately implemented masking filters.