Short name: CoPro5G
Long name: Cooperative proactive resource management for 5G in unlicensed spectrum
Company: Cognitive Innovations Private Company
Call: F4Fp-SME-COD1 (see call details)
Proposal number: F4Fp-SME-COD190122-02
Call Stage 2: F4Fp-SME-2 (see call details)
Proposal number: F4Fp-SME-STAGE 2- 06M23
SUMMARY REMARKS & TESTBEDS
We propose cooperative proactive resource management for 5G in an unlicensed spectrum. Unlicensed spectrum can be exploited by legacy systems such as 5G through licensed assisted access (LAA). LAA will be deployed into multiple small base stations (SBSs) providing a large-scale wireless network. Afterwards, a two-stage machine learning (ML) solution will be employed to the SBSs in order to provide learning of the activity in an unlicensed spectrum. The ML module will be deployed in the multiple 5G SBSs that will be able to aggregate the unlicensed spectrum bands. The ML will rely on a double Q-learning algorithm that will take into account the interference also in such a wireless network deployment.
Given the ML implementation, a proactive resource allocation will be provided taking into account the interference from all sources and learning the activities in advance. Next, cooperation will be also provided among the SBSs by employing in-band communication in the 5G (LTE) bands. Cooperation will enhance the decision to access the unlicensed bands jointly by sharing knowledge about the channel occupation. Experimental results will prove the concept of cooperative proactive resource management in the unlicensed bands compared to a non-cooperative solution.
- CoPro5G experiment – Stage 2 (download the poster)
- Review CoPro5G experiment – Stage 2 (download the slides)