TinyTurbo: Efficient Turbo Decoders on Edge

TL;DR

TinyTurbo introduces a neural-augmented Turbo decoder that achieves near-MAP performance with low complexity, robustness across practical channels, and strong generalization, verified through over-the-air tests.

Contribution

The paper presents TinyTurbo, a neural-enhanced Turbo decoder that outperforms traditional methods in reliability while maintaining low complexity and broad applicability.

Findings

Achieves near-MAP decoding performance with low complexity

Demonstrates robustness on LTE-standard channels like EPA and EVA

Verifies effectiveness through over-the-air experiments

Abstract

In this paper, we introduce a neural-augmented decoder for Turbo codes called TINYTURBO . TINYTURBO has complexity comparable to the classical max-log-MAP algorithm but has much better reliability than the max-log-MAP baseline and performs close to the MAP algorithm. We show that TINYTURBO exhibits strong robustness on a variety of practical channels of interest, such as EPA and EVA channels, which are included in the LTE standards. We also show that TINYTURBO strongly generalizes across different rate, blocklengths, and trellises. We verify the reliability and efficiency of TINYTURBO via over-the-air experiments.