Grazia Fattoruso

Semi-Supervised Learning Techniques in Artificial Olfaction: A Novel Approach to Classification Problems and Drift Counteraction

Semi-supervised learning is a promising research area aiming to develop pattern recognition tools capable to exploit simultaneously the benefits from supervised and unsupervised learning techniques. These can lead to a very efficient usage of the limited number of supervised samples achievable in many artificial olfaction problems like distributed air quality monitoring. We believe it can also be beneficial in addressing another source of limited knowledge we have to face when dealing with real world problems: concept and sensor drifts. In this paper we describe the results of two artificial olfaction investigations that show semi-supervised learning techniques capabilities to boost performance of state-of-the art classifiers and regressors. The use of semi-supervised learning approach resulted in the effective reduction of drift-induced performance degradation in long-term on-field continuous operation of chemical multisensory devices.

An open source GIS system for earthquake early warning and post-event emergency management

The recent advances in geo-informatics have been opening new opportunities in earthquake early warning and emergency management issues. In the last years, the geo-scientific community has recognized the added value of a geo-analytic approach in complex decision making processes for critical situations due to disastrous natural events such as earthquakes. In fact, recently, GIS-based solutions are investigated in several research projects such as SIT_MEW Project, aimed at the development of volcanic and seismic early warning systems (EWSs). In this project context, an innovative open source GIS system has been investigated and developed as integrated component of the seismic EWS. Its architecture consists in a geospatial database system, a local GIS application for analyzing and modelling the seismic event and its impacts and supporting post-event emergency management, a WEB-GIS module for sharing the geo-information among the public and private stakeholders and emergency managers involved in disaster impact assessment and response management.

A maker friendly mobile and social sensing approach to urban air quality monitoring

Novel model of citizenship calls for a new approach to the policy making, characterized by the wish to be part of the information building process. The citizen wants to become an active member of the smart city. This has its impact also on the air quality monitoring and control process. In this work, we try to answer to these needs by investigating a citizen centered air quality monitoring concept. The goal is to enable individuals to monitor their exposure to air pollution and simultaneously to contribute creating a map of the state of urban air quality through the sharing of data.

Collaborative Open Source Geospatial Tools and Maps Supporting the Response Planning to Disastrous Earthquake Events

The latest improvements in geo-informatics offer new opportunities in a wide range of territorial and environmental applications. In this general framework, a relevant issue is represented by earthquake early warning and emergency management. This research work presents the investigation and development of a simple and innovative geospatial methodology and related collaborative open source geospatial tools for predicting and mapping the vulnerability to seismic hazard in order to support the response planning to disastrous events. The proposed geospatial methodology and tools have been integrated into an open source collaborative GIS system, designed and developed as an integrated component of an earthquake early warning and emergency management system.

An Ontology Framework for Flooding Forecasting

Floods can cause significant damage and disruption as they often affect highly urbanized areas. The capability of knowledge using and sharing is the main reason why the ontologies are suited for supporting the phases of forecasting in (near-) real time disastrous flooding events and managing the flooding alert and emergency. This research work develops an ontology, FloodOntology for floods forecasting based on continuous measurements of water parameters gathered in the watersheds and in the sewers and simulation models. Concepts are captured across the main involved domains i.e. hydrological/hydraulic domains and SN-based monitoring domain. Classes hierarchies, properties and semantic constraints are defined related to all involved entities, obtaining a structured and unified knowledge-base on the flooding risk forecasting, to be integrated in expert systems.

An Holistic Approach to e-Nose Response Patterns Analysis—An Application to Nondestructive Tests

Artificial olfaction is an emerging application field for machine learning practitioners. In this paper, we propose a holistic approach to pattern classification in electronic noses applications. In particular, we show how classification results based on a complete measurement cycle can be combined with an assessment provided by real-time classifiers acting on the single instantaneous measurement sample. A running classification confidence measure allows for obtaining fast and reliable outcomes. A safety critical scenario has been selected for the testing of the proposed pattern analysis strategy involving the identification and discrimination of surface contaminants on composite panels in pre-bonding nondestructive tests during lightweight aircraft assembly. A reject option has been introduced to refuse low classification confidence panels improving both FP and FN rates. Results show how this strategy can efficiently exploit two different views of the electronic nose olfactive fingerprinting process that is currently seen as alternative.

7.4.5 Wireless Chemical Sensor Networks for Air Quality Monitoring

Air quality assessment and monitoring could be efficiently performed by a distributed network of cost effective chemical sensors cooperating for the reconstruction of a chemical image of the sensed environment. The obtainable insights are of paramount importance in many applications ranging from city air pollution monitoring to energy efficiency in smart buildings. However several research studies have highlighted the challenging nature of developing such architecture. Actually, the intrinsic properties of the sensed phenomena, those of the commercially available chemical sensors including power requirement and stability, together with the needed communication infrastructure make a real world implementation of such a system a problem worth of significant investigation efforts. This paper encompass the ENEA-UTTP efforts in providing solutions to several challenges both in indoor and outdoor air quality monitoring setups.

SISI Project: Developing GIS-Based Tools for Vulnerability Assessment

In the framework of a wider research project, aimed at developing an high performance computing infrastructures in Southern Italy, the SISI project aims at developing customized GIS tools for vulnerable site selection in Italian territory. Vulnerable sites will be defined based on co-presence of different dangerous/vulnerable geographic features in given areas. Morphological and topological criteria will possibly be included in territorial diagnostics.

A SWE Architecture for Real Time Water Quality Monitoring Capabilities Within Smart Drinking Water and Wastewater Network Solutions

The world is facing a water quantity and quality crisis. These global concerns are addressing water sector operators to smart technological solutions that realize the so-called smart drinking water and wastewater networks. Water quality preservation is one of the essential services that smart water utilities have to guaranteed. The water quality monitoring systems include a variety of in situ sensors with several sensor protocols and interfaces. Sensor integration as well as real time sensor readings accessibility and interoperability across the interconnected layers of functionality needed for a comprehensive smart water network solution are the challenges should be tackled. The objective of this research work has been to develop a standardized OGC SWE (Sensor Web Enablement) architecture that enables the integration and real time access to the various continuous and networked sensors can be installed along drinking water and wastewater networks, and real time sensor data browsing, querying and analyzing capabilities across the components of a smart water network solution. Furthermore, a web based geo-console and a QGIS SOS client application have been developed ad hoc for supporting utilities to effectively manage their water treatment and optimize quality-testing processes.

An adaptive immune based anomaly detection algorithm for smart WSN deployments

The growing attention in smart WSN deployments for monitoring, security and optimization applications urges the design of new tools in order to recognize, as soon as a possible, anomalous states of systems whenever they occur. In order to develop an anomaly detection system enabling to discover unusual events in a non-stationary process, a scalable immune based strategy has been adopted. The algorithm works as an instance based 1-class classifier capable to un-supervisedly model the “normal” spatial-temporal variable behavior of the system identifying first order anomalies. Typical immune-like processes guarantee a slow adaptation of the set of local patterns to long term variation in the monitored system. The algorithm has been applied to a several real scenarios showing to be able to work on both on resource constrained WSN nodes and on dealing with large data streams in centralized data processing facilities.