Forums

Search results: 4

 
img2.png

The activity is within a NATO (North Atlantic Treaty Organization) research project on security for underwater acoustic communications. The student will study techniques for establishing a secret key, i.e., a sequence of bits shared by communications between two devices and not disclosed to an eavesdropping device. The solution will be based on the estimation of the communication channel between the two legitimate devices that will operate as a unique signature of their link, not known to the eavesdropper. Information theoretic tools will be used to tackle the problem. The internship and final project activity will a) investigate the characteristics of the underwater acoustic communication channel that can be used for key agreement, b) design techniques to obtain the secret key, based on information theoretic approaches, and c) simulate the key agreement procedure with suitable tools that describe the behavior of underwater acoustic channel.

Required skills: Basics of probability and signal processing for communications. Matlab and Python programming languages.

Languages (write/speak): Italian or English

Period: Open

Support: 400 €/month for 6 months.

University contact: Prof. Stefano Tomasin (stefano.tomasin@unipd.it)



 
Our position and movements reveal many sensitive information, including our acquaintances, health conditions, work, lifestyle, and buyer preferences. Recently, the localization of smartphones has been considered for contact-tracing in the fight against the COVID-19 outbreak. In this occasion, the privacy of position and movement information was widely debated, and solutions based on ranging rather than positioning were widely adopted, still raising various privacy concerns.

img2.png

In 5G cellular systems, the user position is explicitly known by the network in precision of decimeters and passed also to external applications. Moreover, the transmission of unencrypted control signals by the smartphone allows an eavesdropper to know and track the user position. In cases, a virtual private mobility network (VPMN) could be deployed to ensure privacy.

The concept of VPMN is similar to the conventional virtual private network (VPN) used at the network layer to obtain a private sub-network over the Internet. The VPMN operates across all the communication layers, as it provides that UEs directly communicate among themselves and mutually operate as relays with the cellular networks. Therefore, a message departing from UE-A first directly goes to UE-B (without involving the cellular network) and then it is forwarded to the gNB by UE-B on behalf of UE-A. In this way, transmission of the message from UE-A occurs from UE-B, whose position may be revealed to external eavesdroppers. For Example, consider that VPMN wants to conceal being within a building. By using the VPMN, it first establishes a VPMN connection with UE-B, a device that is in its proximity, but outside of the building. UE-A then exchanges all messages with the cellular network usingUE-B as a proxy, which that has the position information on UE-B.Note that the VPMN does not need a single point of attachment to the cellular network, i.e., all UEs of the VPMN can communicate (in uplink) with the cellular network. The key point is that the transmitting UEdoes not reveal the position from which a message is originated. This is achieved for example by randomly choosing the UE that forwards (in uplink) the packetto the cellular network. For downlink transmission, the cellular network will do an anycast transmission to theVPMN UEs, which will then route the packet to the correct destination UE.

The thesis can proceed in various directions:

- the evaluation of the VPMN ability to hide the location (probability of finding users around, in specific directions, location uncertainty at the base station…);

- the additional costs of delay incurred by packets doing multiple hops, and additional power consumption of collaborating nodes;

- the discovery of nearby nodes and the determination of the routing path
the design of beamforming in order to direct signal towards other nodes of the VPMN, without leaking information to the base station.

Thesis suited for students of ICT for Internet and Multimedia and Computer Science

For further information, contact Prof. Stefano Tomasin (stefano.tomasin@unipd.it)
 
An intelligent reflective surface (IRS) is a surface that reflects radio signals, and its reflective characteristics can be adjusted. Think of a smartphone that does not see the base station since there is a building obstructing the direct path. If a reflective surface is properly positioned between the smartphone and the base station, the signal transmitted by the base station can be reflected by the surface and reach the smartphone, by circumventing the building. Intelligent reflective surfaces are made of many elements that can be adapted in part, in order to better focus the signal to the specific smartphone position. A simple mathematical model of the reflective surface is available.

img1.png

IRSs have emerged in the recent months as a very attractive technology to extend the coverage of cells and provide new features: they combine a very simple structure (thus making them cheap to manufacture, install, and maintain) while providing relevant benefits. Still, many challenges and opportunities are related to IRSs. For possible applications and technical challenges of IRSs, see the link to the paper below.

Various topics are available for a thesis on IRS, including:


- Localization using IRSs: the high focusing capabilities of RISs of large geometric size can be capitalized for finely estimating the location of mobile terminals and devices, so as to support high-precision ranging, radio localization, and mapping applications.

- Beamforming design with multiple IRSs. We can direct signals from the base station to the IRS by using beamforming techniques, that is, using multiple antennas at the base station and transmitting signals with different phases that coherently sum at the intended destination. We can use beamforming also to null the signal in specific positions, thus reducing interference to other receiver. The presence of IRSs in a cell requires new solutions for the design of beamforming, taking into account that we can also properly configure the IRSs in order to focus the signal in specific direction and avoid transmitting in other directions.



For further information, contact Prof. Stefano Tomasin (stefano.tomasin@unipd.it)

 
Smartphones consume a significant amount of energy. There are plenty of cases where simple actions may reduce the smartphone power consumption, including:

- use of WiFi instead of cellular network
- switch off of the radio link when no connection is available (e.g., entering a tunnel)
- buffering data to be transmitted to reduce overhead (send big packets at long interval of time, rather than many small packets frequently)

However, in order to improve user experience and further reduce energy consumption, prediction techniques should be deployed in order to understand when it is time to switch off the radio (or reduce the packet rate) and when we can switch it on again. For example, when alternating between WiFi and cellular networks, we must switch on the cellular link before we exit from the coverage area of WiFi, to avoid service interruptions. Machine learning techniques can be useful, and we must also assess their energy impact in the whole procedure.

The master thesis will investigate the problem of energy reduction in the above-mentioned scenarios (and possibly others), and study the application of simple prediction (machine-learning) techniques to decide the behavior of the smartphone (when to switch radio on and off). An assessment of performance and suitable energy-efficient learning solutions will be pursued.

For this thesis it is possible to do an internship at the laboratory of DEI.
For more information contact Prof. Stefano Tomasin (tomasin@dei.unipd.it).