Authors: Nitu Shah, Saima Khan, Dr. Vikas Gupta, Sandip Nemade
Abstract: This study explores the application of smart algorithms on improved adaptive modulation and coding schemes in 6G systems. The rapid development of wireless communication requires system requirements advanced sufficient to dynamically maximize the spectral efficiency, simultaneously the ultra-reliable low-latency communication. By using machine learning-based intelligent modulation and coding scheme (MCS) selection, such as deep reinforcement learning with a Q-learning method and CNNs, this study attempts to fine-tune MCS selection in 6G networks. The conclusion from the proposed hypothesis is that intelligent algorithm driven adaptive MLC can deliver substantially higher performance in throughput/spectral efficiency/BER as compared to traditional lookup table and outer loop link adaptation methods. Results show that modular adaptive modulation and coding based on reinforcement learning achieves 10%-20% gain in terms of throughput over traditional outer loop link adaptation, and spectral efficiency gains lie between 12.64% ∼ 21.52% for different velocity conditions. The computational complexity of deep learning methodologies decrease up to 80%, with comparable block error rate performance. Results indicate that intelligent algorithms can achieve real-time channel quality adjustment, and improve key 6G performance metrics. This work paves the way for self- organizing wireless networks supporting a variety of quality-of-service demands in future generation communication systems.
DOI: http://doi.org/10.5281/zenodo.18231654
