Showing papers by "Consolatina Liguori published in 2015"

ISO/IEC/IEEE 21451 Smart Sensor Network for the Evaluation of Motorcycle Suspension Systems

Abstract: Smart sensors and transducers are being widely employed in the development of electronically controlled systems, such as advanced suspensions, to improve the performance guaranteed by terrestrial vehicles. Nevertheless, a standard reference for the full exploitation of the smart features is not yet extensively adopted. The CAN-bus is only used for the communication functionalities within a sensor network, resulting in cost-effective and more reliable connections. An original data-driven approach, based on both the CAN-bus and the IEEE ISO/IEC/IEEE 21451 standard, is proposed for the characterization of suspension systems of two-wheel vehicles with regard to the comfort and vehicle handling. Because of the ISO/IEC/IEEE 21451, the data acquired from sensor are sent by means of a GPS-General Packet Radio Service module to PC software that implements the processing algorithm. The proposed methodology includes a low-cost measurement setup, data processing algorithms in both time and frequency domain, and the introduction of model-free synthetic indices to experimentally compare the on-road behavior influenced by different damping characteristics for the suspension system. The method is applied to the on-road verification of the riding comfort assured by different control strategies for semiactive suspensions.

Towards the evaluation of the measurement uncertainty of environmental acoustic noise

Abstract: It is observed that the difference between a measurable value and a threshold does not involve a straightforward comparison of values, given that such measurement, which can only be an estimation, is expressed as an interval. Consequently, the value has to be considered to possess a clear degree of uncertainty. Such matters are highly relevant to the area of the measurement of relative levels of environmental noise, from the perspective of providing an adequate estimate of the indetermination associated with such measurements. Here it is necessary to account for the uncertainty that appears around particular noise events. The intention of the first phase of this study is to eliminate outliers that occur when measuring signals in real time using a statistical approach based on histograms. Then the second phase will determine the uncertainty associated with the measurement of ‘purified’ signals.

Towards the Evaluation of the Measurement Uncertainty of Environmental Acoustic Noise Outlier Detection

Abstract: It is observed that the difference between a measurable value and a threshold does not involve a straightforward comparison of values, given that such measurement, which can only be an estimation, is expressed as an interval. Consequently, the value has to be considered to possess a clear degree of uncertainty. Such matters are highly relevant to the area of the measurement of relative levels of environmental noise, from the perspective of providing an adequate estimate of the indetermination associated with such measurements. Here it is necessary to account for the uncertainty that appears around particular noise events. The intention of the first phase of this study is to eliminate outliers that occur when measuring signals in real time using a statistical approach based on histograms. Then the second phase will determine the uncertainty associated with the measurement of 'purified' signals.

Towards visual smart metering exploiting wM-Bus and DLMS/COSEM

Abstract: Automatic Meter Reading (AMR) is one of the key issues in modern energy efficiency technologies, which are expected to bring on great benefits to both customers and utility companies (electricity, gas and water). Several efforts have been recently made in developing suitable communication protocols for metering data management and transmission. In order to match the cost requirements for the water market, an original solution represented by a visual sensor is proposed as an add-on module for the smart meter. The adoption of developed electronic device and ad-hoc protocol at application layer within the last mile of a multiservice Advanced Metering Infrastructure is investigated in terms of energy issues and interoperability. This study also analyses the communication performance that can be expected when using wireless M-Bus technology to send DLMS/COSEM messages and complete an image file transmission from the smart meter to data concentrator unit.

A Proposal for the Management of the Measurement Uncertainty in Classification and Recognition Problems

Abstract: This paper proposes a new approach to classification and recognition problems. It considers the measurement uncertainty affecting input data to improve the overall effectiveness of this type of process. The proposed method is based on an effective probabilistic approach for the evaluation of the confidence level of system outputs and the suitable use of related information to improve the performance in terms of the correct decision rate. As a case study, it is applied to a particular face recognition classification algorithm based on linear discrimination analysis. The performance comparison with a traditional approach has proven the value of the proposal.

A Software Sensor for Motorcycle Suspension Stroke

Abstract: An increasing adoption of sensors is demanded for implementing dynamics control systems in many contexts and in particular in the motorcycle field, where strong constraints on costs, safety and bulks have also to be met. This paper is about the estimation of suspension stroke exploiting the analytical redundancy among the measurement signals provided by a typically adopted instrumentation set. A software sensor for the rear suspension is designed according to a systematic approach, which is focused on recurrent Artificial Neural Networks able to predict the dynamic behavior of the vehicle. Experimental results show the rear suspension elongation can be correctly estimated. They also disclose the possibility of setting-up an effective Instrument Fault Detection and Isolation scheme based on the real-time adoption of the proposed software sensor in order to improve the system reliability.

Smart meters and demand response: A double challenge

Abstract: The evolution of energy demand and production can be faced with a Grid that becomes Smart: a network that is capable of integrating heterogeneous players in both generation and demand, allowing each of them to access electricity market and being able to react in real-time and autonomously to changing conditions on the grid. Such paradigm may allow to provide advanced services (i.e. Demand Response) increasing network efficiency and reliability. Necessary requisite for the implementation of a Smart Grid is having a metering architecture that can provide timely and reliable information regarding the state of the grid. The paper addresses the topic of an Advaced Metering Infrastructure (AMI), focusing on the first block, the Smart Meter. A meter prototype for the electric energy measurement, based on ARM architecture, is presented; the technical solutions, both in hardware and software, are described in detail; finally the results of tests carried out for its metrological characterization are reported.

A web-based application for dermoscopic measurements and learning

Abstract: Image processing techniques have been long proposed for automatic analysis of skin lesions in the field of early detection of melanoma. Nevertheless, Computer Aided Systems are not yet able to outperform the diagnostic accuracy of expert dermatologists. They could instead reveal very useful in providing with a second opinion and improving the detection results from physicians with short clinical experience. The paper introduces an original web-based application for the automatic detection of dermoscopic structures within pigmented lesions and the support to novel dermatologists. It is able to receive and store digital images captured by digital cameras and smartphones equipped with dermoscopy, measure morphological and chromatic parameters, and take into account the measurement uncertainty to finally provide a clinical decision according to the diagnostic method 7-Point Checklist.

Machine vision systems for on line quality monitoring in industrial applications

Abstract: This paper is concerned with the topic of non-contact measurements for industrial applications. It presents the case of vision-based systems realized by the authors expressly for the measurement of geometric and/or chromatic parameters of rubber profiles for the automotive industry. After a brief description of the extrusion process, a stereo vision system for the on-line measurement of the dimensional characteristics of the profile transversal section is first described in all its main components and features, then two modules for the inspection of the finish of profile surfaces are presented.

Semi-active RFID Devices for Traceability

Abstract: In this paper a semi-active RFID module for remote monitoring of storage conditions during the transportation is proposed. It integrates one SmartTag and some sensors, and allows also to locate the object to the one which is related. In order to obtain the real time observing on the transportation vehicle an On Board Unit (OBU) is installed that works as a concentrator and communicates with a central control room.

The Use of Uncertainty for Improving the Reliability of Classification in Face Based Recognition Algorithms

Abstract: The paper proposes a new approach to classification and recognition problems which takes into account the measurement uncertainty affecting input data for improving the overall reliability of such kind of processes. The proposed method is based on an effective probabilistic approach for the evaluation of the confidence level of system outputs and the suitable use of such information for improving the performance in terms of correct decision.

A proposal for improving the performance of face recognition systems based on 3d features

Abstract: In this paper a suitable methodology for the improvement of the reliability of results in classification systems based on 3D images is proposed. More in detail, it is based on the knowledge of the uncertainty of the features constituting the 3D image (obtained processing a pair of two 2D stereoscopic images) and on a suitable statistical approach providing a confidence level to the classification result. These pieces of information are then managed in order to improve the classification performance in terms of correct classification and missed classification percentages. The experimental results, obtained applying the methodology on an Active Appearance Models algorithm, a popular method for face recognition based on 3D features, show that, compared with a traditional approach (which generally does not take into account the uncertainty on 3D features), the proposed methodology allows to significantly improve the classification performance even in scenarios characterized by a high uncertainty.