Network management and control
We investigate computational intelligence in the context of networks, with particular focus on the Internet of Things, Software Defined Networks, and Opportunistic Communications.
The group is well known for its work on learning-based optical, wireless and sensor networks. It investigates how networks can make sense of complex factors (usage patterns, traffic shapes, interference, energy availability etc.) and self-regulate, just like the networks of Nature have evolved to do.
The research focuses on three main strands: swarm intelligence (i.e. how pursue global properties through simple local mechanisms); autonomic computing (i.e. how to obtain a coordinated response in face of extreme perturbations); and cognition (i.e. how to adapt to unpredicted conditions, rewire, self-regulate and evolve).
The Future Internet will have to handle the massive variety of signals generated by trillions of nodes. It will have sensing and actuation capabilities. It will face unprecedented levels of interference, dynamics, power consumption, and unpredictability. It will be asked to deliver speed but also stability and reliability.
To address these challenges, the Smart Communication Networks group aims to bring data mining and human cognition concepts directly into the networks, seeking new ways to realize distributed machine learning mechanisms, lightweight in-node learning, and cloud-assisted inference.