Short name: Fed4AI (stage1 & stage2)
Long name: Experimental validation of AI models for the prediction of wireless links performance

Country: Portugal

Call: F4Fp-SME-COD1 (see call details)
Proposal number: F4Fp-SME-COD1-09

Call Stage 2: F4Fp-SME-2 (see call details)
Proposal number: F4Fp-SME-STAGE 2- 06M11


Accurate prediction of wireless performance links can be very useful to optimize radio resource allocations. However, the vast variety of possible wireless configurations and propagation scenarios make it notoriously hard to design explicit\theoretical models to forecast the performance of wireless links. They exhibit several intricacies due to complex interactions between the PHY and MAC layers, which manifest themselves in frequency, spatial and time domains. Existing radio link performance models usually adopt explicit and bottom-up approaches in order to predict throughput figures based on Markov chains and SINR levels. In this experiment, we would like to carry out an experimental validation of a different approach for predicting the performance of radio links for Wi-Fi (Phase 1) and LTE (Phase 2). Rather than manually fitting analytical models to capture complex dependencies, we are going to directly learn the models themselves, using artificial intelligence and machine learning techniques with a limited set of observed measurements. The integration of these advanced algorithms with ALLBESMART’s network analytics framework UXPERT, constitutes an undeniable competitive advantage over our competitors, transforming UXPERT from reactive to proactive with significative added value and business impact.


Start typing and press Enter to search