Abstract

We present trellis codes designed to perform well in the context of periodic symbol puncturing. Progressive puncturing of these new trellis codes allows unequal error protection or continuous adaptation of the information rate to match channel and traffic conditions. Deep fades or strong interference also may be considered as a form of puncturing. For these environments, we present examples of trellis codes that are robust to all puncturing patterns that maintain non-negative effective redundancy. This robustness extends even to puncturing patterns that reduce the effective redundancy to zero. 1. Introduction Although no previous work is available specifically on puncturing a trellis code, the related problem of puncturing convolutional codes (sometimes as part of a trellis code design) has been treated extensively. Cain, Clark, and Geist [2] obtained good high rate convolutional codes by periodically puncturing a lower rate code. Using this approach, Yasuda, Kashiki, and Hirata [15]...

Citations