MAB in OSA (2015)

Presented at Dyspan 2015, CROWNCOM 2016

This demonstration presents a proof-of-concept for opportunistic spectrum access. It particularly focuses on reinforcement learning algorithm called UCB (Upper Confidence Bound) designed by the machine learning community to solve the MAB problem (Multi-Armed Bandit). The demonstrator shows the first worldwide implementation of reinforcement learning algorithms for OSA (opportunistic spectrum access) on real radio environment using USRP N210 platforms.

Video

 

For more information, you can read the following article:

C. Moy, A. Nafkha and M. Naoues, “Reinforcement learning demonstrator for opportunistic spectrum access on real radio signals,” 2015 IEEE International Symposium on Dynamic Spectrum Access Networks (DySPAN), Stockholm, 2015, pp. 283-284.