Networking Protocols & Signal Processing, Architectures, Body Sensor Networks. In the past decade we have been studying a broad range of protocols for Internet of Things from data gathering, authentication / encryption  to wireless reprogramming and network control. Our latest research addresses the IoT architecture, the spatio-temporal compression of data from various application domains. Our current activities include energy harvesting sensor networks, joint design of in-network signal processing and communication and body sensor networks.


The term Internet of Things describes a number of technologies and research disciplines that will enable the Internet to reach out into the real world of physical objects. Technologies like RFID, short-range wireless communications, real-time localization, ad hoc and wireless sensor networks (WSNs) are now becoming increasingly common and all will take part in the Internet of Things (IOT). According to the IOT paradigm, physical objects will be equipped with some communication capabilities, which will be exploited to coordinate their action and in particular the way these objects influence the surrounding physical space. A common example is that of the books in a library, which could be equipped with RFID tags, so that each book could be precisely located by a WSN system deployed in the library. This information could then be fed to any search engine running on a computer located within the library or even outside (given that this computer has the right access credential for this data). This example could be easily extended to our houses or offices, so that it will be possible to acquire the physical location of objects within buildings through a dedicated WSN infrastructure connecting the physical world to the Internet domain. A user could then “google” for a book using standard Internet browsers that, inside the home or work network, will be empowered with IOT services. The same rationale holds in the case IOT networks perform actions on the physical space such as locking doors, turning on video cameras for surveillance, controlling modern lighting/heating/cooling systems for optimal comfort, while minimizing the energy expenditure.

Architectural design

This line of research deals with the interconnection of WSNs with the Internet space. At the beginning of the WSN era, researchers focused on the development of dedicated systems, where highly specialized but nonstandard protocols were used within the WSN, whereas one or multiple gateways were used to translate messages and ultimately connect the WSN to the external IP world. While these systems were generally efficient in the specific application scenario they were designed for, they lacked flexibility: developing new applications on top of them was therefore time-consuming and cumbersome, as it required modifications to the specialized protocols within the WSN. Recently, the 6LoWPAN standard has been proposed as a viable method to bring IPv6 to WSNs, so that sensor nodes can be natively addressed and connected through the IP protocol. This has obvious advantages such as rapid connectivity and compatibility with preexisting architectures, plug-and-play installation of WSNs, rapid development of applications as well as the possibility of integrating with existing Web services developed for standard IP networks. Recent research efforts explore the feasibility and the performance limits of tiny Web services (also called binary Web servers) on top of 6LoWPAN for WSNs. Other issues are related to using compressed XML for communicating with binary Web servers.

Networking protocols design

We are concerned with the design of energy efficient, scalable and fully autonomous IoT systems. This entails the design of joint in-node and in-networks compression algorithms along with collection and recovery systems that minimize the amount of information that is collected from large deployments (thus minimizing the amount of data traffic generated within the network). All of this shall be carried out while 1) minimizing the energy consumption per useful bit and 2) delivering the monitored signals with sufficient accuracy at the application layer. Other issues are related to the wireless reprogramming/reconfiguration/retasking of large WSNs, which should be done in an autonomous and secure manner. In addition, a number of problems are related to the interconnection of heterogeneous devices (in terms of transmission, memory and processing capabilities) and to the utilization of energy scavenging techniques within WSNs (e.g., solar powered devices or wireless transfer). The latter, in particular, affects job scheduling and channel access operations; it is still unclear how energy scavenging can be optimally used within distributed and multi-hop systems featuring in-network processing (e.g., spatio-temporal compression, data aggregation, data fusion). Ongoing work includes:

  • In-node and in-network processing: joint design of processing and communication (channel access, routing);
  • spatio-temporal compression: including the use of compressive sensing, data aggregation, data fusion;
  • energy harvesting systems: joint design of communication and scheduling / control in the presence of harvested ambient energy or wireless energy transfer.

Wearable devices and body sensor networks

We are particularly interested in biomedical applications and in particular in the realtime processing of vitals from wearable IoT devices. We target chest bands, smart watches or wristbands, which permit the collection of biometrics data (e.g., heart-rate, oxygen level, respiration, blood pressure, etc.) that can be used to help address the individual health and fitness needs of the users. Our present research activity deals with the design of a lightweight and integrated software for on-device and real-time processing of vital signals, so that they can be effectively processed, stored in the limited memory of wearable IoT devices and conveniently transmitted over their wireless interface. We want the system to adapt to the signals being sampled, by being prompt when required by the application and gently go into some power saving mode when the signals exhibit regular patterns. Note that the system should be ready to snap back to the full operating mode as soon as new activity is detected. Also, high resolution should be provided when the user is up to some dynamic activity and wants to track that or when a critical behavior is detected. Ongoing work includes:

  • runtime compression of biometric time series for their efficient transmission over wireless channels;
  • motion analysis: gait analysis from smartphone and wristband motion data, user identification and pattern analysis for preventive healthcare applications;
  • activity recognition and behavioral profiling from motion data (gyroscope and accelerometer).

Related projects

  • INTEL ISRA EC-CENTRIC (2015-2018)
  • GRO SAMSUNG AWARD (biomedical data processing for IoT devices, 2014-2015)
  • FP7 EU SWAP (energy harvesting sensor networks, 2010-2014)
  • FP7 EU IoT-A (Internet of Things architectures, 2010-2013)
  • CARIPARO WISE-WAI (City-wide intelligence through Wireless Sensor Networking, 2008-2010)


  • NEC Europe Ltd
  • University of Twente
  • CTTC: Centre Tecnològic de Telecomunicacions de Catalunya
  • University College of Dublin
  • Docomo Eurolabs
  • Intel