Pietro Siciliano

Support vector machines for olfactory signals recognition

Pattern recognition techniques have widely been used in the context of odor recognition. The recognition of mixtures and simple odors as separate clusters is an untractable problem with some of the classical supervised methods. Recently, a new paradigm has been introduced in which the detection problem can be seen as a learning from examples problem. In this paper, we investigate odor recognition in this new perspective and in particular by using a novel learning scheme known as support vector machines (SVM) which guarantees high generalization ability on the test set. We illustrate the basics of the theory of SVM and show its performance in comparison with radial basis network and the error backpropagation training method. The leave-one-out procedure has been used for all classifiers, in order to finding the near-optimal SVM parameter and both to reduce the generalization error and to avoid outliers.

On the study of feature extraction methods for an electronic nose

In this study, we analyzed the transient of microsensors based on tin oxide sol–gel thin film. A novel method to this research field for data analysis and discrimination among different volatile organic compounds is presented. Moreover; several feature extraction methods have been considered, both steady-state (fractional change, relative, difference and log) and transient (Fourier and wavelet descriptors, integral and derivatives) information. Feature extraction methods have been validated qualitatively (by using principal component analysis) and quantitatively on the classification rate (by using a radial basis function neural network). # 2002 Elsevier Science B.V. All rights reserved.

Preparation, characterisation and applications of thin films for gas sensors prepared by cheap chemical method

Abstract The physical properties of pure and doped (Ni, Os, Pd and Pt) SnO 2 thin films, prepared by using a chloride-based inorganic sol–gel route, have been reported. These properties were investigated by using differential scanning calorimetry and thermogravimetric (DSC/TG), FTIR, X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD) and atomic force microscopy (AFM) techniques. Microsensors for gas detection were also fabricated and tested in various atmospheres (nitrogen oxide, carbon monoxide, ethanol, methanol and methane). The dopants affected the coating properties depending on the formation of nanoparticles. Furthermore the gas-sensing properties towards reducing and oxidising gases were found to be dependent on the nature of doping element. An array consisting of the manufactured sensors has been arranged and examples of applications for “electronic nose” are given.

Synthesis, electrical characterization, and gas sensing properties of molybdenum oxide nanorods

Gas sensors have been prepared by using single crystalline MoO3 nanorods. Structures 50–150nm and up to 15μm in length were synthesized by template directed hydrothermal synthesis. Gas sensing tests showed responses to NO2 up to 740% and to NH3 up to 570%. Moreover, preliminary electrical characterizations were performed, as a function of temperature in 20–520°C range temperature, by means of current-voltage measurements. I-V dependence studies show a linear increase in current at low voltage and a quadratic increase at higher voltages. A study of the current-temperature dependence shows that at high temperature current is thermally active and successively decreases that can be explained by taking in account a change of material morphology.

An active vision system for fall detection and posture recognition in elderly healthcare

The paper presents an active vision system for the automatic detection of falls and the recognition of several postures for elderly homecare applications. A wall-mounted Time-Of-Flight camera provides accurate measurements of the acquired scene in all illumination conditions, allowing the reliable detection of critical events. Preliminarily, an off-line calibration procedure estimates the external camera parameters automatically without landmarks, calibration patterns or user intervention. The calibration procedure searches for different planes in the scene selecting the one that accomplishes the floor plane constraints. Subsequently, the moving regions are detected in real-time by applying a Bayesian segmentation to the whole 3D points cloud. The distance of the 3D human centroid from the floor plane is evaluated by using the previously defined calibration parameters and the corresponding trend is used as feature in a thresholding-based clustering for fall detection. The fall detection shows high performances in terms of efficiency and reliability on a large real dataset in which almost one half of events are falls acquired in different conditions. The posture recognition is carried out by using both the 3D human centroid distance from the floor plane and the orientation of the body spine estimated by applying a topological approach to the range images. Experimental results on synthetic data validate the correctness of the proposed posture recognition approach.

Monitoring of rancidity of milk by means of an electronic nose and a dynamic PCA analysis

Abstract This paper reports the use of a semiconductor thin films based electronic nose to recognise the rancidity of two different kinds of milk (UHT and pasteurised) during their ageing days. The employed sensor array consists of five different SnO 2 thin films prepared by means of the sol–gel technology. The data coming from the response of the sensors have been elaborated by principal component analysis (PCA) in order to obtain a classification of the data clusters related to different milk ageing days and so track the dynamic evolution of milk rancidity. The obtained results could lead to industrial applications in dairying for the quality control of starting milk and the rancidity check during the different stages of production.

Physical characterization of hafnium oxide thin films and their application as gas sensing devices

Thin films of hafnium oxide have been prepared by using a dual ion beam sputtering system. A study of their physical properties is reported. In particular, structural and compositional characterization was performed by means of x-ray diffraction and x-ray photoelectron spectroscopy techniques, showing a mixture of amorphous and polycrystalline structure and a substoichiometric composition. The atomic force microscopy results have shown a crater-like morphology probably due to the deposition process. In addition the gas sensing characteristics were analyzed in the presence of carbon monoxide. The variations in the electrical resistance have shown the capability of the films to detect CO and then the possibility to use hafnium oxides as a new sensitive material in the field of gas sensors.

Analysis of vapours and foods by means of an electronic nose based on a sol–gel metal oxide sensors array

Abstract In this paper, a novel electronic nose based on a metal oxide gas sensors array, prepared by means of the sol–gel method in the form of thin films, is described, including the fabrication process. Two different configurations of the array were developed: one consisting of sensors with the same sensing layer (SnO2) but different geometry and material for the electrical electrodes, and another consisting of chemically modified SnO2 thin films. Examples of application to the analysis of air pollutants (CO, NO2, CH4, C2H5OH and CH3OH) and different foods (oil, milk, tomato, wine, ground coffee). Recognition of different air contaminants and of the aroma of foods is achieved by using the principal component analysis (PCA). Good discrimination among the clusters has been obtained indicating promising results for the application of our sensor array as electronic nose, in particular in the food industry for the classification of different olive oils.

A novel gas sensor based on SnO2/Os thin film for the detection of methane at low temperature

Abstract Osmium-doped tin oxide thin films were prepared by means of sol–gel technique. Their methane sensing properties were studied and compared to undoped SnO 2 thin films. The good sensitivity to methane at a low working temperature makes them very attracting active materials to be used in domestic CH 4 leak-detectors.

Properties of vanadium oxide thin films for ethanol sensor

Thin films of vanadium oxides for ethanol sensor applications were been prepared by means of rf reactive sputtering from a vanadium pentoxide target in an Ar–O2 atmosphere. Films with the best sensitivity and selectivity were obtained with 15% oxygen at a working temperature between 280 and 300 °C. These films show a very low sensitivity to CO, CO2, and CH4 and total insensivity to NH3.