Nicola Saccomanno

Nicola Saccomanno


Supervisor: Angelo Montanari

+39 0432 558457

Stanza / Room: L2-06-ND (DMIF, NS3)

Research Project

Advanced solutions for indoor positioning

Indoor positioning has become a quite relevant topic over the last few years thanks to its application in fields like detection of objects (stored goods, medical equipment, etc.), detection of people either during everyday activities or in the context of emergencies, personalised information delivery, healthcare monitoring, context awareness, and so on.

Nowadays, several solutions are currently being exploited to address this problem. One of the most effective is fingerprinting. A core advantage behind the adoption of this technique is that it relies only on WLANs, which are widely deployed today. Indoor fingerprint-based positioning solutions are characterised by two steps. The first one, called offline (or training) phase, aims at building the radio map used in the second phase. Basically, at certain predefined reference points (i.e., specific locations of the building), multiple readings of the Wi-Fi network are sampled and stored into a database. The second, online, phase occurs when a device is interested in identifying its position. To achieve that, it submits its observation of the Wi-Fi network, which is subsequently compared with those stored in the database to determine the possible location of the device. The literature identifies some challenges related to classical fingerprint-based solutions as well as more advanced ones. Some examples are the management of devices heterogeneity, the deployment optimisation, the radio map construction, the handling of faulty measurements, the low accuracy of classical solutions, the access points selection, the lack of uniformity of the signals, and the absence of universal standards for both indoor positioning and indoor mapping.

During my PhD, I am addressing the indoor positioning topic and related issues from a data science perspective. Precisely, by exploiting different techniques ranging from machine learning to more classical ones, I am interested into investigating how to take advantage of data to refine the current state of the art and possibly contributing in solving some of the current challenges. Among the possible lines of research, I am planning to study how to combine indoor and outdoor positioning, how to identify system accuracy at runtime, how to address radio map construction major related issues, and how the introduction of more refined buildings maps can improve the current scenario.