MALIN – Multi-Armed Bandit Learning in Iot Networks (2018)

This demo was presented at ICT 2018 by Lilian Besson, Rémi Bonnefoi and Christophe Moy.

In this demo, we implement an IoT network the following way: one gateway, one or several intelligent (learning) objects, embedding the proposed solution, and a traffic generator that emulates radio interferences from many other objects. Intelligent objects communicate with the gateway with a wireless ALOHA-based protocol with no specific overhead for learning needs. We model the network access as a discrete sequential decision making, and using the framework and algorithms from Multi-Armed Bandit (MAB) learning, we show that intelligent objects can improve their access to the network by using load complexity and decentralized algorithms, such as UCB and Thompson Sampling.

More information about the poster is given in this poster:
http://www-scee.rennes.supelec.fr/wp/wp-content/uploads/2018/07/MALIN_poster.pdf

We made a short video which describes this demonstration:

You can also read the paper:

Bonnefoi, R.; Besson, L.; Moy, C.; Kaufmann, E.; Palicot, J. “Multi-Armed Bandit Learning in IoT Networks: Learning helps even in non-stationary settings”, CROWNCOM, September 2017.

The open-source GNU-Radio code used for this demo is available at:
https://bitbucket.org/scee_ietr/malin-multi-arm-bandit-learning-for-iot-networks-with-grc