5G • 5G-Advanced (Rel-18)
AI/ML Link Adaptation
Deep Learning Predictive Modulation and Coding Scheme (MCS) Selection
Uses deep neural networks to dynamically predict channel variations and assign optimal MCS levels, outperforming legacy Outer Loop Link Adaptation (OLLA).
Technical Explanation
Legacy Outer Loop Link Adaptation (OLLA) adjusts SINR offsets reactively based on ACK/NACK history, which lags during fast fading or bursty traffic. Rel-18 AI link adaptation analyzes channel history, velocity, and interference patterns to predict instantaneous SINR, choosing the highest possible MCS that meets block error rate targets without wasting spectral efficiency.
Key Functions
- Replaces slow reactive OLLA algorithms with proactive neural channel prediction
- Maximizes spectral efficiency by operating close to instantaneous Shannon capacity
- Reduces BLER spikes during sudden acceleration and high Doppler shifts
- Supports fast AMC (Adaptive Modulation and Coding) up to 1024-QAM
Specifications
3GPP TR 38.843, TS 38.214 (Rel-18)
Interfaces
Uu (DCI Format 1_1/1_2)
Related 5G Concepts
Want to memorize 5G concepts like this one?
Study it with SuperMemo SM-2 spaced repetition flashcards.