Irina E. Bocharova

BEAST decoding of block codes obtained via convolutional codes

BEAST is a bidirectional efficient algorithm for searching trees. In this correspondence, BEAST is extended to maximum-likelihood (ML) decoding of block codes obtained via convolutional codes. First it is shown by simulations that the decoding complexity of BEAST is significantly less than that of the Viterbi algorithm. Then asymptotic upper bounds on the BEAST decoding complexity for three important ensembles of codes are derived. They verify BEASTs high efficiency compared to other algorithms. For high rates, the new asymptotic bound for the best ensemble is in fact better than previously known bounds.

On the covering radius of convolutional codes

We consider a problem of calculating covering capabilities for convolutional codes. An upper bound on covering radius for convolutional code is obtained by random coding arguments. The estimates on covering radius for some codes with small constraint length are presented.

On Trellis Codes for Linear Predictive Speech Compression

We consider the problem of efficient codebook search for the code excited linear predictive (CELP) speech coders. To dimmish the search complexity a trellis codebook is proposed and decoding algorithms based on the Viterbi algorithm are investigated. Numerical results of the computer simulation are presented.

Fast Infinite Response Filtering for Speech Processing

The problem of reducing the number of multiplications in infinite response filtering is considered. We show how this operation may be decomposed on finite response filtrations and so-named halfconvolutions. Examples of the efficient procedures for filters of low order are presented.