Andrea L Piroddi

In the area of low-power wireless networks, one technology that many researchers are focusing on relates to positioning methods such as fingerprinting in densely populated urban areas. This work presents an experimental study aimed at quantifying mean location estimation error in populated areas. Using a dataset provided by the University of Antwerp, a neural network was implemented with the aim of providing end-device location. In this way, we were able to measure the mean localization error in areas of high urban density. The results obtained show a deviation of less than 150 m in locating the end device. This offset can be decreased up to a few meters, provided that there is a greater density of nodes per square meter. This result could enable Internet of Things (IoT) applications to use fingerprinting in place of energy-consuming alternatives.

The aim of this research is to explore a possible method for identifying an alphanumeric character produced by an individual with motor and vocal disabilities and whose possibilities to interact with the surrounding environment are limited to the movement of the face. The reason behind the creation of an alphanumeric character interpretation algorithm is linked to the fact that attempts to use existing algorithms (such as Tesseract) gave poor results in this scenario. The reason of the low success rate was not investigated; the focus was on finding a functional solution to the objective. Obviously subsequent studies in this sense are desirable. A Human-Computer Interaction algorithm has been developed in order to allow a person with limited mobility to interact with a computer, to which commands can be given.  This project is based on three main elements: the identification of facial parameters, their tracking and the interpretation of the word expressed. A Linux based Python applic...

One of the most complex challenges that wireless communication systems will face in the coming years is the management of the radio resource. In the next years, the growth of mobile devices, forecast (CISCO, 2020), will lead to the coexistence of about 8.8 billion mobile devices with a growing trend for the following years. This scenario makes the reuse of the radio resource particularly critical, which for its part will not undergo significant changes in terms of bandwidth availability. One of the biggest problems to be faced will be to identify solutions that optimize its use. This work shows how a combined approach of a Reinforcement Learning model and a Supervised Learning model (Multi-Layer Perceptron) can provide good performance in the prediction of the channel behavior and on the overall performance of the transmission chain, even for Cognitive Radio with limited computational power, such as NB-IoT, LoRaWan, Sigfox.

The aim of this research is to explore a possible method for identifying an alphanumeric character produced
by an individual with motor and vocal disabilities and whose possibilities to interact with the surrounding
environment are limited to the movement of the face. The reason behind the creation of an alphanumeric
character interpretation algorithm is linked to the fact that attempts to use existing algorithms (such as
Tesseract) gave poor results in this scenario. The reason of the low success rate was not investigated; the
focus was on finding a functional solution to the objective. Obviously subsequent studies in this sense are
desirable. A Human-Computer Interaction algorithm has been developed in order to allow a person with
limited mobility to interact with a computer, to which commands can be given. This project is based on three
main elements: the identification of facial parameters, their tracking and the interpretation of the word
expressed. A Linux based Python application has been coded to elaborate information from a camera. The
main action is to identify the face of the potential user, activate tracking when the subject closes his eyes for a
defined numbers of times, then follow the trajectory of a point of the face (in our case the tip of the nose),
interpret the letter thus drawn, buffer the letters and finally convey the command (entire word) thus obtained.
At the end, statistics on the success percentage are provided

— The article investigates the possibilities of applying machine learning algorithm to identify an individual through biometric voice recognition with the higher possible reliability. The emphasis in the analysis is placed on the possibility of using artificial intelligence approach methods for the purposes of recognizing a person unambiguously, uniquely on the basis of the data contained in his/her vocal spectral information. A large number of routes we can go to parametrically representing the speech signal for the voice recognition system such as Mel-Frequency Cepstrum Coefficients (MFCC). During the authentication phase the input voice signal is recorded and processed comparing it by using MFCC features with a signal that has been previously stored in the database by the same user. The main purpose is to compare some of the main machine learning algorithms to classify them on this particular application

—Wireless communication is in its nature volatile due to rapidly varying propagation conditions between the transmitters and receivers. The propagation conditions inside the tunnels are different from what is experienced in areas outside the tunnels. This paper presents a new method to calculate received power in a defined point inside the tunnel.

—This paper presents an analysis of the behavior of a fluid control system in the presence of turbulence, and a method to reduce the measurement error, due to the non-linearity of the surface profile. The algorithm that measures the distance of the fluid surface from the chosen point, drives an ultrasonic sensor which is rather sensitive to the fluctuations of the outer layer of the fluid, so it is therefore important to find a way to avoid any feedback control which is not linked to a real dangerous situation.

— The object of this study is to analyze the ramp profile functions of buried asymmetric dielectric targets for non-axial or oblique incidence of the electromagnetic wave. This approach can be used, with due limitations, in satellite SAR radar to scan large areas of land on which it intends to identify buried metallic or dielectric objects (such as landmines, etc..). The results contained in this dissertation are based on (a scattering computation using a Method of Moments (MoM) solutions with SCILAB simulation tool. The targets used in the simulation include metallic and dielectric spheres and rectangular metallic and dielectric blocks.