S. De Vito

Wireless Sensor Networks for Distributed Chemical Sensing: Addressing Power Consumption Limits With On-Board Intelligence

Chemicals detection and quantification is extremely important for ensuring safety and security in multiple application domains like smart environments, building automation, etc. Characteristics of chemical signal propagation make single point of measure approach mostly inefficient. Distributed chemical sensing with wireless platforms may be the key for reconstructing chemical images of sensed environment but its development is currently hampered by technological limits on solid-state sensors power management. We present the implementation of power saving sensor censoring strategies on a novel wireless electronic nose platform specifically designed for cooperative chemical sensing and based on TinyOS. An on-board sensor fusion component complements its software architecture with the capability of locally estimate air quality and chemicals concentrations. Each node is hence capable to decide the informative content of sampled data extending the operative lifespan of the entire network. Actual power savings are modeled and estimated with a measurement approach in experimental scenarios.

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.

TinyNose: Developing a wireless e-nose platform for distributed air quality monitoring applications

In this work we present the development and proof of concept testing of a protoype wireless e-nose (w-nose) architecture capable of mesh shaped networking. The proposed w-nose is based on a TelosB by Crossbow Inc. and custom, power aware, TinyOS based components for data gathering and local processing. Sensor nodes are equipped with a small array of nonconductive polymer/CB based chemiresistors operating at room temperature for VOCs indoor monitoring. A properly developed conditioning stage board connects the sensor array to the microcontroller ADC. A single w-nose has been tested in a controlled test chamber for terpenes discrimination, while networked motes operation have been demonstrated in ad-hoc small testing facilities for acetic acid spill detection.

Predictive models for building's energy consumption: An Artificial Neural Network (ANN) approach

Buildings energy demand is influenced by many factors, such as: weather conditions, building structure and characteristics, energy consumption of components (lighting and HVAC systems), level of occupancy and users behavior. As consequence of multi-variable impact on buildings energy consumption, theoretical models based on first principles are not able to forecast actual energy demand of a generic building. In this paper, an Artificial Neural Network (ANN) model applied to a real case consisting in a dataset of monthly historical building electric energy consumption is developed. Results show that accuracy of energy consumption forecast runs, in terms of RMSPE (root mean square percentage error), in the range 15.7% to 17.97% and varies slightly according to the prediction horizon (3 months, 6 months and 12 months).

Calibrating chemical multisensory devices for real world applications: An in-depth comparison of quantitative machine learning approaches

Abstract Chemical multisensor devices need calibration algorithms to estimate gas concentrations. Their possible adoption as indicative air quality measurements devices poses new challenges due to the need to operate in continuous monitoring modes in uncontrolled environments. Several issues, including slow dynamics, continue to affect their real world performances. At the same time, the need for estimating pollutant concentrations on board the devices, especially for wearables and IoT deployments, is becoming highly desirable. In this framework, several calibration approaches have been proposed and tested on a variety of proprietary devices and datasets; still, no thorough comparison is available to researchers. This work attempts a benchmarking of the most promising calibration algorithms according to recent literature with a focus on machine learning approaches. We test the techniques against absolute and dynamic performances, generalization capabilities and computational/storage needs using three different datasets sharing continuous monitoring operation methodology. Our results can guide researchers and engineers in the choice of optimal strategy. They show that non-linear multivariate techniques yield reproducible results, outperforming linear approaches. Specifically, the Support Vector Regression method consistently shows good performances in all the considered scenarios. We highlight the enhanced suitability of shallow neural networks in a trade-off between performance and computational/storage needs. We confirm, on a much wider basis, the advantages of dynamic approaches with respect to static ones that only rely on instantaneous sensor array response. The latter have been shown to be best choice whenever prompt and precise response is needed.

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.

Dynamic multivariate regression for on-field calibration of high speed air quality chemical multi-sensor systems

In the last few years, the interest in the development of new pervasive or mobile implementations of air quality multi-sensor devices, has significantly grown. New application opportunities appeared together with new challenges due to limitations in dealing with rapid pollutants concentrations transients. In this work, we propose a Dynamic Neural Network (DNN) approach to the stochastic prediction of air pollutants concentrations by means of chemical multi-sensor devices. DNN architectures have been devised and tested in order to tackle the cross sensitivities issues and sensors inherent dynamic limitations. Testing have been performed using an on-field recorded dataset from a pervasive deployment in Cambridge (UK), encompassing several weeks. Results have been compared with those obtainable by static models showing the performance advantages of on-field dynamic multivariate calibration in a real world air quality monitoring scenario.

Simulation of Chlorine Decay in Drinking Water Distribution Systems: Case Study of Santa Sofia Network (Southern Italy)

Drinking water treatment is needed for providing water that is safe from disease-causing pathogenic microorganisms. Chlorine is widely used as a disinfectant in drinking water systems, although the main disadvantages are the decay of its concentration along the pipes and the formation of undesirable by-products (DBPs). In this research work, the chlorine decay process has been investigated along a real water distribution system, the Santa Sofia aqueduct (Campania, Southern Italy), by an innovative modeling approach. The adopted hydraulic and water quality models have been calibrated on live data (the physical/chemical characteristics of the drinking water) and gathered continuously by a wireless network of multi-parametric probes distributed along the Santa Sofia aqueduct. The residual chlorine concentrations throughout the Santa Sofia network, predicted by hydraulic and water quality models calibrated on the same aqueduct, may be considered reliable.

Use of Kinetic Models for Predicting DBP Formation in Water Supply Systems

Drinking water chlorination reduces the risk of pathogenic infection, but it may be harmful to human health because of disinfection by-product (DBP) formation. Available predictive models of DBP formation are almost exclusively calibrated at lab scale. The objective of the present research work is to apply two of them at full scale for the Santa Sofia aqueduct (Campania, Southern Italy), in order to predict DBP formation and evolution as function of the real water network characteristics. Live data, gathered continuously by a wireless network of multi-parametric probes, installed on the aqueduct, along with data measured in laboratory, are used for model calibration. The predictive scenarios are performed by using an open source integrated GIS-based platform (including Epanet, MSX, GIS uDig).

ENABLING DISTRIBUTED VOC SENSING APPLICATIONS: TOWARD TINYNOSE, A POLYMERIC WIRELESS E-NOSE

In this work, we present the development of a novel wireless e-nose platform designed for indoor distributed VOC detection and quantification. The proposed w-nose, called TinyNose, rely on a small polymeric sensor array that is connected to a commercial wireless mote by means of custom developed electronics. A custom developed software architecture allow for signal acquisition, processing and transmission to a data sink where data are stored and/or presented to the remote user. In this work a preliminary assessment of TinyNose capabilities to operate in open air configuration is conducted by using different source of indoor VOC pollution to be detected and classified by the developed architecture.