Umberto Maniscalco

A virtual layer of measure based on soft sensors

In this paper it is proposed a method to design and train a layer of soft sensors based on neural networks in order to constitute a virtual layer of measure in a wireless sensor network. Each soft sensor of the layer esteems the missing values of some hardware sensors by using the values obtained from some other sensors. In so doing, we perform a spatial forecasting. The correlation analysis for all parameter taken into account is used to define a cluster of real sensors used as sources of measure to esteem missing values. An application concerning the fire prevention field is used as a test case and results evaluation.

Analysis and visualization of meteorological emergencies

The capability to sample and store meteorological information across a wide area allows to analyze the historical evolution of data and to extract events that are potentially bound to emergency and critical events. In this contribution we detect events when a station shows values that are sensibly different from the neighbor stations. We check the co-occurrence of these events with emergency reported in web news. Results are encouraging and show how the statistical analysis can allow to forecast emergencies and to reduce the impact of critical situations.

Virtual Sensors to Support the Monitoring of Cultural Heritage Damage

The present work is part of a wider research activity carried on within the Italian National Project named SIINDA. It shows how physical atmosphere parameters like temperature, humidity, wind direction, can be indirectly estimated in specific points of the monument, if one, or more than one, ambient air monitoring station is present in the neighborhood of the monument itself. We use a connectionist system trained to map the parameters measured by such stations with the parameters measured by the set of installed sensors. The obtained results look like very good and we received the approving by cultural heritage experts who evaluated such a methodology to effective by support monitoring in the field of the conservation state of monuments.

Disaster Prevention Virtual Advisors Through Soft Sensor Paradigm

In this paper we illustrate the architecture of an intelligent advisor agent aimed at limiting, or as far as possible preventing, the damages caused by catastrophic events, such as floods and landslides. The agent models the domain and makes forecasting by exploiting both ontology models and belief network models. Furthermore, it uses a monitoring network to recommend preventive measures and giving alerts, if necessary, before that the event happens. The monitoring network can be implemented through both physical and soft sensors: this choice makes the measurements more adequate and available also in case of failure of some of the physical sensors. The front-end of the agent is made by a chat-bot, capable to interact with human users using natural language.

Soft sensors for social sensing in cultural heritage.

This paper presents a work in progress concerning a soft sensor approach for social sensing in the context of cultural heritage. The approach analyzes public posts and comments on the British Museum Facebook page and tries to give an overall measurement regarding the sentiment and the emotions arising from a post. This can help museums to better address their resources in order to improve the effectiveness of their divulgation action.

Detection of Indoor Actions Through Probabilistic Induction Model

In the present work a system able to classify the indoor action is presented. The data are recorded with multiple kind of sensor collecting the position of the joints of the person in the room, the acceleration recorded on the person wrist and the presence or absence in a specific room. The latent semantic analysis, based on the principal component search, allows to estimate the probability of a given action according the sampled values.

Classification of Indoor Actions through Deep Neural Networks

The raising number of elderly people urges the research of systems able to monitor and support people inside their domestic environment. An automatic system capturing data about the position of a person in the house, through accelerometers and RGBd cameras can monitor the person activities and produce outputs associating the movements to a given tasks or predicting the set of activities that will be executes. We considered, for the task the classification of the activities a Deep Convolutional Neural Network. We compared two different deep network and analyzed their outputs.

Soft Sensor Network for Environmental Monitoring

This paper shows the application of a soft sensor network for the detection of meteorological events. A set of hard (real) sensor are placed in a territory, where they measure heterogeneous quantities. Starting from their measurements, a soft sensor network provides useful information coming from the data. In this contribution we show how prediction and validation of data can be done through machine learning approach by collecting data from the historical series. Furthermore, we show how the cluster based on correlation analysis among the data achieved by the sensors can be sensibly different from the ones simply drawn on geographical distance.

A fully visual based business document classification system

A fully visual approach for business documents classification is presented. The paper describes how SURF visual features, extracted from the documents, can be usefully used for business document recognition and their classification. Some of the extracted features are used to compute a prototype aiming at speed up the comparison of a document class while obtaining the best recognition rate. Moreover, we can determine which features are relevant and we can select zones of interest in the documents. Experimental setup has been performed on a set of real business documents of different typologies and companies. We tested also the robustness of our approach adding artificial defects and noise to the original documents and classifying them taking into account exclusively visual and graphical features. The capability of documents classification without any kind of text analysis has the great advantage to make the system totally independent from the idiom.

Adaptive Semantics of Complex Information/Services Networks: A Case Study from Cultural Heritage

A unified view to Semantic Web/Grid Information and Services Discovery and Management is being investigated at CNR both at the theoretical level and at the experimental level. The mathematical foundations are being developed with algebraic methods. An experimental prototype is being realized in the area of Cultural Heritage. The paper presents achievements and future directions.