Mixed Block Markov Superposition Transmission Codes
2026-06-15 • Information Theory
Information Theory
AI summaryⓘ
The authors study a type of error-correcting codes called BMST codes, which help transmit data reliably. They note that two existing versions each have trade-offs: one avoids persistent errors but can cause error spreading, while the other tends to have error floors but no propagation. To improve this, the authors combine both types in new mixed BMST codes that take advantage of their strengths. Their simulations show these new codes perform better with less memory needed than the recursive version alone.
Block Markov superposition transmissionerror-correcting codesrecursive codesnon-recursive codeserror propagationerror floorconcatenationsliding-window decodingframe error ratebit error rate
Authors
Philipp Mohr, Jasper Brüggmann, Viet Hoang Le, Gerhard Bauch
Abstract
Block Markov superposition transmission (BMST) codes provide a flexible framework for constructing codes with near-capacity performance and low-complexity sliding-window decoding. However, existing BMST variants show contrasting performance limitations: recursive BMST (rBMST) codes suffer from error propagation but avoid high error floors, whereas non-recursive BMST codes exhibit the opposite behavior. Motivated by these complementary characteristics, we combine recursive and non-recursive components through parallel and serial concatenation, yielding mixed BMST (mBMST) codes. The proposed framework subsumes existing BMST variants and enables new BMST structures. Simulations show that these structures improve FER and BER performance with lower memory requirements than rBMST.