|
ИСТИНА |
Войти в систему Регистрация |
ИСТИНА ЦЭМИ РАН |
||
An Algorithm Is Described for Predicting the Probability of Success of Signal Transmission in a Wireless Communication System Using Machine Learningстатья
- Автор: Bobrov E.A.
- Журнал: Moscow University Computational Mathematics and Cybernetics
- Том: 46
- Номер: 3
- Год издания: 2022
- Издательство: Allerton Press Inc.
- Местоположение издательства: United States
- Первая страница: 117
- Последняя страница: 124
- DOI: 10.3103/s0278641922030037
- Аннотация: A dynamic machine learning algorithm is described for predicting the probability of successful signal transmission and adaptive signal modulation and coding in a massive MIMO (Multiple Input Multiple Output) system. The algorithm is based on a fully connected neural network, which is initially trained on the output of the standard OLLA (Outer Loop Link Adaptation) algorithm, and then gradually trained on the basis of feedback from the wireless system. Numerical modeling of the MIMO system is done in different scenarios to confirm the quality of the proposed algorithm, using different characteristics of the communication channel and speeds of the served users. The advantage is shown of the proposed algorithm over the modern Q-learning algorithm and the standard OLLA scheme.
- Добавил в систему: Бобров Евгений Александрович