Massimo Carota

Towards the investigation of kinematic parameters from an integrated measurement unit for the classification of the rising from the chair

In this paper we introduce a new method to evaluate the ability to rise from a chair by means of the sit-to-stand locomotion task. It is based on the analysis of the vertical acceleration peaks and the timing as assessed by an our designed device. Preliminary results indicate the feasibility of discriminating the rising from a chair fixed to different heights and the discrimination by pathological and non pathological Parkinsonian subjects

A new integrated kinematic sensor for the classification of sit-to-stand locomotion task

In this paper we introduce a new kinematic sensor to evaluate the ability to rise from a chair by means of the sit-to-stand locomotion task. It is based on the analysis of the acceleration assessed by a homemade accelerometric transducer. Preliminary results show the feasibility of discriminating the rising from a chair fixed to different heights and the capability of distinguishing between pathological and non pathological parkinsonian subjects.

A cellular neural network based character recognition system

In this paper we present a character recognition system based on cellular neural networks. As a consequence, we deal with a real-time system. All considered features can be extracted by CNN templates; recognition is merely feature-based, with no need of a learning phase or any kind of memory.

Mental task recognition based on SVM classification

In this paper, a brain/computer interface is proposed. The aim of this work is the recognition of the will of a human being, without the need of detecting the movement of any muscle. Disabled people could take, of course, most important advantages from this kind of sensor system, but it could also be useful in many other situations where arms and legs could not be used or a brain-computer interface is required to give commands. In order to achieve the above results, a prerequisite has been that of developing a system capable of recognizing and classifying four kind of tasks: thinking to move the right hand, thinking to move the left hand, performing a simple mathematical operation, and thinking to a carol. The data set exploited in the training and test phase of the system has been acquired by means of 61 electrodes and it is formed by time series subsequently transformed to the frequency domain, in order to obtain the power spectrum. For every electrode we have 128 frequency channels. The classification algorithm that we used is the Support Vector Machine (SVM).

Discrimination between human functional ability/disability by means of different classification methodologies

A new method for discovering well defined incipient pathologies in human beings, by means of the analysis of the sit- to-stand locomotion task, is proposed. It is based on the frequency analysis of acceleration measurements supplied by a homemade transducer and on the exploitation of some of the most effective classification strategies at this time. Results show the capability of distinguishing between pathological and non pathological subjects.

A new physical sensor based on neural network for musical expressivity

In this paper, we present an innovative physical sensor interface based on neural network that allows an electronic music composer to plan and conduct the musical expressivity of a performer. For musical expressivity we mean all those execution techniques and modalities that a performer has to follow in order to satisfy common musical aesthetics. The proposed sensor interface is composed by a gestural transducer, that measure motion acceleration and angular velocity, and a mapping module, that transform few physical measured parameters into a lot of specific sound synthesis parameters. It is able to transform six physical input parameters in seventeen sound synthesis parameters. In this work, we focus our attention on mapping strategies based on Neural Network to solve the problem of electronic music expressivity.

A customizable tool for cellular nonlinear network simulation

In this paper we introduce a new release of the CELL cellular neural network simulator. It has been extended with new capabilities, which allow users, among other things, to add custom state equations or new learning algorithms, in order to simulate a larger set of non-linear neural networks.

Fully asynchronous neural paradigm

In this paper, a new integrate-and-fire neural model is proposed. In order to obtain an asynchronous behaviour, a different delay time is assigned to every firing process, in function of the inner dynamics of the single neuron. This model allows also an efficient computer simulation, which has been implemented with MATLAB and supports up to a hundred of thousands of neurons. Simulations show an auto-confinement property: when proper input is given, we observe a specific neural group selection in which the activity appears in a quite high level, while it remains lower in the regions among different groups. When the considered input is terminated, activity seems to remain stable, acting as a memory of some past input event.

The Sky-Scanner System for Air Traffic Management: Test Sessions and Statistical Analysis

Air Traffic Management (ATM) is traditionally performed by means of radar systems, but new detection and ranging systems based on LIDAR (LIgth Detection And Ranging systems) are emerging as a critical design trend and yielding to new generation ATM paradigms. The main goal of the designers of the system under discussion was to develop a novel laser tracking technology (SKY-Scanner System) that could allow the detection and tracking of aircrafts up to at least 6 nautical miles. Moreover, the purpose was also the definition of techniques, protocols, numerical prediction tools and devices specifically designed for the analysis of the laser system performances in Air Traffic Control applications. The ultimate purpose was to define a new generation ATM paradigm that is based on radar and laser tracking data fusion, as well as on ground to air laser communications. This paper reports the test session that has been run on a prototype of the laser range finder, in order to find accuracy and possible dependence of measurement errors on temperature, humidity, distance and position of the target. The results show that high humidity and temperature can affect the accuracy of the measurements.

Improving Piano Music Transcription by Elman Dynamic Neural Networks

In this paper, we present two methods based on neural networks for the automatic transcription of polyphonic piano music. The input to these methods consists in live piano music acquired by a microphone, while the pitch of all the notes in the corresponding score forms the output. The aim of this work is to compare the accuracy achieved using a feed-forward neural network, such as the MLP (MultiLayer Perceptron), with that supplied by a recurrent neural network, such as the ENN (Elman Neural Network). Signal processing techniques based on the CQT (Constant-Q Transform) are used in order to create a time-frequency representation of the input signals. The processing phases involve non-negative matrix factorization (NMF) for onset detection. Since large scale tests were required, the whole process (synthesis of audio data generated starting from MIDI files, comparison of the results with the original score) has been automated. Test, validation and training sets have been generated with reference to three different musical styles respectively represented by J. S. Bach’s inventions, F. Chopin’s nocturnes and C. Debussy’s preludes.