On error exponents for woven convolutional codes with outer warp and unequal error protection

Abstract

Woven convolutional codes are a family of error forward correction codes. An important construction within this code family are woven convolutional codes with outer warp. They allow us to provide unequal error protection for the encoded information. The performance of woven convolutional codes with outer warp and unequal error protection is studied in terms of error exponents. Two constructions are considered. The first is shown to allow a small part of the information to be essentially better protected without degrading the performance of the other part. The second construction possesses a two-level error protection for a given partitioning of the information. The error exponents for both partitions are given and it is shown how the better protected information can be used in the process of decoding to improve the performance of the less-protected information. Copyright © 2003 AEI.