Internet of Things

Computational intelligence and network science for the internet of things

Everything is being interconnected. Data bigger than ever are to be generated and transferred. Innovative algorithms are being devised to make connections and sense out of seemingly uncorrelated information. DSC/e sets to give to Internet of Things what nature gave to life on our planet, evolution. We strive for a global autonomous sustainable and self-orchestrated ecosystem of everything.

Scope

Connectivity is, already now, possible between virtually anything. Internet of things relies on a multitude of technologies to help anything communicate and share its own data. However, it is the local computational intelligence that will allow smarter things generate more valuable and actionable data. At the intersection of artificial intelligence, network science and internet of things, DSC/e strives to build things able to autonomously create their own code of conduct, self-organize and interact with their environment.   

Our belief:
A global sustainable IoT consumes the data it produces 

Our drivers:

  • Connect anything anywhere anytime
    dependable networks, network science, swarm intelligence, self-organization, security
  • Sense everything, act selectively
    Computational intelligence, data analytics, machine learning, pattern recognition, self-awareness
  • Actionable data is business
    Business models and strategies, standardization, intellectual property

Vision

Autonomous evolution, self-orchestration and global interconnection of everything are features of real ecosystems. An ecosystem of IoT devices is capable of self-organizing in clusters, competing for resources with each other, gaining new capabilities and devising new strategies to boost their own survivability. In such a system, analyzing data is the key to achieving awareness of the local environment and detecting internal and external threats. The interaction of such an ecosystem with the human society, business models and applications will define a dynamic set of rules of operation. 

The realization of IoT systems at scale is a daunting task, involving the digitization of sensing, actuation, and control. Future IoT systems won’t be driven by the deterministic mechanisms of today. They will be sustained by all sorts of intelligent processes, both inside the devices and among them. Things will have the ability to seamlessly interplay but also to survive external perturbations, minimize their digital footprint, self-diagnose, and generally operate based on predictions.

Research challenges

The roadmap of this vision relies on several paths, currently at different maturity levels. 

Business Modeling

  • Redesign data intensive applications and business models and strategies with IoT and Data Science •Valorization and intellectual property strategies despite recent explosion of industry standards 

Deployment

  • Digitize physical world via innovative sensors and data pre-processing and dissemination methods •Design future-proof autonomous and dependable networks ready to deal with massive scale and traffic, complexity and security challenges 

Making Sense

  • Build data mining methods and tools for semantic interoperability of IoT systems and for (near-)real-time complex events detection and correlation •Design methods for automated detection of emergent patterns in data streams

Control Loops

  • Translate patterns in data to learned and preferred actions via online machine learning, prediction, and control •Design dependable control loops to provide end-to-end performance predictability and guarantees 

IoT Ecology

  • Architect global self-evolving IoT systems to attain swarming, competing and behavior emergence characteristics

Risk management

  • Develop methods to manage risk in distributed and interconnected IoT systems, including threat analysis, threat propagation, and strategic response •Design security measures and procedures that effectively represent risk and provide immediate security advice to customers and network managers

Project examples

  • Inter-IoT EU Horizon 2020
    IoT interoperability and data mining to enable any-to-any standard-agnostic communication.
  • ACCUS EU Artemis
    Cooperative control in smart cities breaks silos and allows for dependable survivable networks of devices.
  • DEMANES EU Artemis
    Monitoring and operation in industrial IoT systems fuels the future logistics with high value actionable data.
  • SCOTT EU Horizon 2020 ECSEL
    Secure COnnected Trustable Things to realize dependable data dissemination and control.
  • ProHeal EU ITEA
    Automated self-protection and self-healing software solutions towards a fully autonomous evolving IoT.

Facilities & Datasets

We maintain an experimental open IoT facility, that allows the deployment of applications, cloud-assisted event and data logging and correlation,  IoT data mining, dataset generation and analysis.