| Popis: |
Deliverable D4.4 extends the work conducted within D4.2 “Intelligent D-Band wireless systems and networks initial designs” and includes the final results of the work performed within Task 4.2 “Machine Learning based network intelligence” of WP4 “D-Band wireless network optimization leveraging ML principles”. Within ARIADNE, optimization of D-band network performance has been investigated through different perspectives, i.e., hardware imperfections at the Tx and the Rx, directional connectivity, LOS and NLOS connectivity, prediction of channel parameters, optimization of RISs performance, resource allocation with Users – AP association, mobility management. Optimization has been achieved by applying suitable ML algorithms depending on the problem to be solved. Specifically, our approaches include Reinforcement Learning (RL) and Deep Reinforcement Learning (DRL), hybrid Metaheuristic-Machine Learning models, complex event forecasting, combination of Recurrent Neural Network (RNN), Convolutional Neural Network (CNN) and Constrained Deep Reinforcement Learning (C-DRL) models. Overall, the deliverable analyzes and evaluates innovative methods and concepts towards beyond 5G and 6G communication systems |
