Learning Joint Synchronization, Channel Estimation, and Decoding: Comparison with a Finite-Blocklength Achievability Bound

arXiv:2610.41580 · eess.SP · Submitted 2026-10-30 · Read on arXiv

eess.SP

Submitted: 2026-10-30

Updated: 2026-10-30

Comments: 6 pages

Code: https://github.com/chenbin234/Turbo-Autoencoder-vs.-Finite-Blocklength-Bound

Project page: https://nvlabs.github.io/sionna

License: http://creativecommons.org/licenses/by/4.0/

The gist: Learned schemes that jointly perform tasks such as synchronization and decoding have achieved promising results for short-packet communication.

Terminology

Abstract

Learned schemes that jointly perform tasks such as synchronization and decoding have achieved promising results for short-packet communication. However, it remains unclear how close their performance is to information-theoretic bounds. To investigate this question, we extend the turbo autoencoder (TurboAE) proposed in Dörner et al. (2023) to a single-input single-output waveform channel with an unknown delay and memoryless block fading, for which an achievability bound from finite-blocklength information theory is available. Our extension incorporates a pilot-free, model-guided neural synchronizer and enables joint synchronization, channel estimation, and decoding. At a blocklength of 64 complex symbols and a rate of one bit per channel use, the proposed TurboAE achieves a lower bit error rate (BER) than all evaluated pilot-assisted 5G baselines, with gains of 1.9 dB at a BER of 10-3 over a baseline involving a low-density parity-check (LDPC) code and 1.7 dB over a baseline involving a polar code and successive cancellation list decoding (list size 8). In terms of block error rate (BLER), the TurboAE outperforms the LDPC-coded baseline, but lags behind the polar-coded baseline by about 0.8 dB at a BLER of 10-3. Moreover, all evaluated schemes remain at least 2.1 dB away from a pilot-assisted achievability bound, indicating substantial room for improved designs.

Sources

Related papers