Research Topics

Learning and communications co-design in  next generation wireless networks

The activity considers the integration of advanced AI methodologies within beyond cooperative wireless networks (beyond 5G, B5G). Distributed ML techniques, including federated learning (FL), represent a mushrooming multidisciplinary research area weaving in sensing, communication and learning. FL enables continual model training in distributed wireless systems: rather than fusing raw data samples at a centralized server, FL leverages a cooperative fusion approach where networked, collaborative agents, connected via reliable and low-latency wireless links, act as distributed learners that periodically exchange their locally trained model parameters. The activity explores opportunities of FL in B5G wireless networks.

Wireless Sensing: exploiting mmWaves for a New Sensing Environment

 

Our expertise is structured around future localization and sensing opportunities for beyond fifth generation (B5G) wireless communication systems. In contrast to 5G and earlier generations, in B5G localization and sensing will be built-in from the outset to both cope with specific applications and use cases, and to support flexible and seamless connectivity. The research targets signal processing, machine learning and networking design aspects for passive sensing using transmitters of opportunity, massive MIMO radar and communications convergence in the mmWave band.

Smart RadioVision systems: 

The research activity aims at integrating a new generation of environmental recognition capabilities into embedded IoT radio devices. Radio signals encode a 3D view of all moving/fixed objects traversed by their propagation, and such view can be recovered at the receiver(s) side by data processing and analytic tools. The project develops the key enabling technologies to decode this visual information by large-scale processing of radio signals.

Radio Vision technology is based on the continuous monitoring of wireless signals, it is designed to perform real-time tracking of object motion (humans, robots, machines) in line-of-sight, non-line-of-sight, and through-the-wall scenarios. Subjects affect the electromagnetic field in a predictable way such that it is possible to track their movements in the space. The subjects may be actively avoiding localization, or they may be passive and not emitting any useful signals; however, a wireless network deployed to locate them may contain elements which actively transmit.

For more information please contact (stefano.savazzi@cnr.it) or click here.

Cooperative localization methods for 5G wireless systems

Indoor deployment

Vehicular network deployment (IoV)

This activity focuses on localization technologies for Internet of Things (IoT) and Internet of Vehicles (IoV) scenarios. Both scenarios envision the deployment of dense networks of devices interconnected by device-to-device (D2D) communications, with decentralized sensing and processing capabilities to meet strict time constraints as in Industrial IoT and in safety-critical IoV applications. These networks call for the adoption of distributed architectures that ensure interoperability, scalability, flexibility and robustness. Network localization is essential to support location-based services and assisted/automated driving functionalities. Distributed techniques have become a key approach in this context as they enable devices to augment the localization availability and accuracy – particularly in harsh propagation scenarios such as indoor or bad urban areas – without the support of any remote unit or external infrastructure. Consensus algorithms have been proposed to let the devices self-disclose the location information through D2D cooperation and data sharing. This approach was shown to improve both coverage and accuracy of localization in dense network deployments.

For more information please contact (monica.nicoli@polimi.it)

Deployment and design of Internet-of-Things platforms for oil&gas industry (industrial IoT):

Industrial Wireless Networks systems

Large-scale adoption of dense wireless network technologies in industrial plants is mandatorily paired with the development of methods and tools for connectivity prediction and deployment validation. Layout design procedures must be able to certify the quality (or reliability) of network information flow in industrial scenarios characterized by harsh propagation environments. In addition, these must account for possibly coexisting heterogeneous radio access technologies as part of the industrial internet of things (iIoT) paradigm, easily allow postlayout validation steps, and be integrated by industry standard CAD-based planning systems. The goal of the activity is to set the fundamentals for comprehensive industry-standard methods and procedures supporting plant designer during wireless coverage prediction, virtual network deployment and post-layout verification. The proposed methods carry out the prediction of radio signal coverage considering typical industrial environments characterized by highly dense building blockage. They also provide a design framework to properly deploy the wireless infrastructure in interference-limited radio access scenarios. In addition, the model can be effectively used to certify the quality of machine type communication by considering also imperfect descriptions of the network layout. In the pase years the design procedures have been corroborated by experimental measurements in oil refinery sites (Saipem-eni EST refinery, PetroEcuador Esmeraldas refinery) using industry standard IEEE 802.15.4e, ISA IEC 62734 and WirelessHART systems (IEC 62591) operating at 2.4GHz.

For more information please contact (stefano.savazzi@cnr.it, umberto.spagnolini@polimi.it)

EM Body models for opportunistic WiFi radars :

The research will develop a deep understanding of the effects of body motions on EM fields, as well as novel predictive models ready for standardization. The goal is to advance EM tools/models for real-time prediction of shadowing/fading effects over the whole EM spectrum, as key elements for the evaluation of human motion, intentions and activities. In addition to conventional RF bands adopted for WiFi radios and IoT communication tasks, emerging mm-wave radio communication technologies are also addressed defined over the, yet unexplored, E- and sub-THz bands. These are foreseen in next generation communication standards (5G and beyond).

For more information please contact (vitttorio.rampa@cnr.itgianguido.gentili@polimi.it)