D. Capriglione

A Methodology To Test Instrument Software: An Application To the Diagnostic in Automotive Systems

Aim of this paper is to describe an original methodology for instrument software testing. A black-box approach is proposed to verify if the functional requirements of the software under test are fulfilled. In order to grant the reliability of the method, an experiment robust design technique is also adopted for the generation of the software input test sets. As an example of application the proposed methodology is applied to instrument diagnostic software for automotive systems. The results of validation tests carried out on data coming from a car engine measurement system, are reported and analyzed

Analysis of different wavelet segmentation methods for frequency-domain energy detection based spectrum sensing

Usually energy-detection based spectrum sensing techniques for Cognitive Radios need, as preliminary operation, the spectrum segmentation in smaller sub-bands, in order to apply test statistic over each sub-band rather than over the entire span, resulting in a sensitivity and selectivity improvement of the output. Despite uniform subdivision is the easiest way, the problem resides on the optimal sub-band width to achieve sensing purposes. In this paper, three different approaches based on wavelet transform are proposed as preliminary stage of a frequency domain energy detection scheme. The advantage in using a wavelet based segmentation is: (i) spectrum is segmented according to its irregularities, (ii) segmentation is not uniform but adaptive to the scenario, (iii) the number of sub-bands is usually less than in the uniform case, (iv) it employs a mathematical tool with fewer DoFs than usual optimization procedures and (v) it can be made completely automatic. As confirmed by obtained results, the considered approaches guarantee different levels of accuracy, even if all of them provide good results for the analyzed figures of merit.

Sensitivity analysis of influence quantities on signal-to-noise ratio in face-based recognition systems

Nowadays, face recognition systems are going to widespread in many fields of application, from automatic user login for financial activities and access to restricted areas, to surveillance for improving security in airports and railway stations, to cite a few. In such scenarios, several architectures based on both 2D image analysis and 3D reconstruction are investigated and proposed in literature. The actual performance of such systems in terms of correct decision rate is affected by several quantities of influence mainly concerning the conditions of acquisition of the image to be processed. As an example, the image luminosity, the lens defocus and the movement of a subject during the image acquisition can be sources of uncertainty which propagate up to the final classification result, thus affecting the reliability of a subject identification. In previous papers, the authors proposed suitable uncertainty models for both 2D and 3D based architectures able to quantify on-line the level of confidence to assign to the output of such systems according to the ISO-GUM. The proposed models required, for each quantity of influence, to estimate separately their deviations with respect to the reference values achieved in ideal acquisition conditions during the training phase. On the other hand, the quality of an image may be linked to the more general concept of signal-to-noise ratio (SNR), because noise affects the pixel of the image, thus introducing uncertainty on the final image. Therefore, looking for the development of a more straight and simple to use uncertainty model, in this paper the relationships among the quantities of influence and the image SNR are investigated. This activity represents the first step toward the realization of face-based recognition systems able to assign a level of confidence to the output results starting only from the evaluation of SNR on the input image.

How to perform fully compliant spectrum sensing in IEEE 802.22 framework

The paper proposes a customized frequency domain energy detection spectrum sensing scheme. It is needed to fulfill requirements provided by the IEEE 802.22 standard, aimed at extending Internet services to Television White Space frequencies. It provides a detection scheme allowing to achieve strict sensitivity required by the protocol either in sensing time or in detection capabilities. In particular, considered constraints in time and computational burden are tailored to make the method effective in low cost Software Defined Radios, that are envisioned to be typical devices working into IEEE 802.22 framework. Such method should guarantee stability and repeatability in its performance. Therefore, the authors propose to combine standard ROC curve analysis to measurement accuracy evaluation, to return the optimal work-point.

Experimental analysis of software network emulators in packet delay emulation

Network emulators are powerful tools that allow to test and characterize network equipment and protocols. The market offers several measurement solutions. The most widely adopted are software network emulators and among them surely free or open-source emulators are the most used especially in research contexts. Unfortunately, software network emulators are not distributed with technical support and certified performance information. Certified information, about the ability of these software to accurately reproduce the wanted network conditions, is very important to validate the results obtained during design activity and test campaigns that involve them. In this paper, an experimental analysis of the ability of two widely used network emulators to emulate several imposed network conditions in different test set-ups is carried out. In particular, we have focused the attention on analyzing the fidelity in the generation of both static and random packet delays under different operating conditions. Obtained results confirm that imposed conditions are not equally reproduced and particular attention has to be paid in specific situations, as very low static delay imposed conditions.