Publications Lists     [ all BibTeX ]  

In the context of activity recognition, wearable devices arenowadays the preferable hardware thanks to their usability, user expe-rience and performances; at the same time, these devices present limi-tations in terms of computational capability and memory, which forcethe algorithm design to be at the same time ecient and simple. Inthis work, we adopt Symbolic Aggregate Approximation (SAX), a sym-bolic approach for information retrieval in time series data that allowsdimensionality and numerosity reduction; SAX is employed here, in com-bination with 1-Nearest Neighbor classier, to identify activity phases incontinuous repetitive activities from inertial time-series data. The pro-posed approach is validated on a public activity recognition dataset.

The problem of daylight estimation in a smart lighting system is considered. The smart lighting system consists of multiple luminaires with collocated occupancy and light sensors. Using sensor information, the objective is to attain illumination levels higher than specified values at the workspaces. We consider a training phase wherein light sensors are used at the workspaces in addition. Data from the light sensors at the ceiling and workspaces is used to estimate the mapping across the sensors. In the operational phase, the estimated mapping is used at the lighting controller to obtain an estimate of the illuminance value at the workspaces. Under the constraint that the estimated illuminance is higher than a specified target value, the controller optimizes the dimming levels of the luminaires to minimize power consumption. We evaluate the performance of the proposed approach in an open-office lighting model by considering different daylight conditions.

The Multirate Support feature has been introduced by the IEEE 802.11 standard to improve system performance. It has been widely exploited within general purpose Wireless LANs by means of Rate Adaptation (RA) strategies, that unfortunately revealed ineffective for the case of real-time industrial communications. This paper presents the innovative Rate Selection for Industrial Networks (RSIN) algorithm, specifically conceived for the real-time industrial scenario with the goal of minimizing the transmission error probability, while taking into account the deadlines imposed to packet delivery.

In this paper an approach to deal with Predictive Maintenance (PdM) problems with time-series data is discussed. PdM is a important approach to tackle maintenance and it is gaining an increasing attention in advanced manufacturing to minimize scrap materials, downtime, and associated costs. PdM approaches are generally based on Machine Learning tools that require the availability of historical process and maintenance data. Given the exponential growth in data logging in modern equipment, time series dataset are increasingly available in PdM applications. To exploit time series data for PdM, a functional learning methodology, namely Supervised Aggregative Feature Extraction (SAFE), is here employed on a semiconductor manufacturing maintenance problem.

We address the problem of distributed unconstrained convex optimization under separability assumptions, i.e., the framework where a network of agents, each endowed with local private multidimensional convex cost and subject to communication constraints, wants to collaborate to compute the minimizer of the sum of the local costs. We propose a design methodology that combines average consensus algorithms and separation of time-scales ideas. This strategy is proven, under suitable hypotheses, to be globally convergent to the true minimizer. Intuitively, the procedure lets the agents distributedly compute and sequentially update an approximated Newton-Raphson direction by means of suitable average consensus ratios. We show with numerical simulations that the speed of convergence of this strategy is comparable with alternative optimization strategies such as the Alternating Direction Method of Multipliers. Finally, we propose some alternative strategies which trade-off communication and computational requirements with convergence speed.

The problem of emergent synchronization pat-terns in a complex network of coupled oscillators has caughtscientists’ interest in a lot of different disciplines. In particular,from a biological point of view, considerable attention has beenrecently devoted to the study of the human brain as a networkof different cortical regions that show coherent activity duringresting-state. In literature, there can be found different large-scale models of resting-state dynamics in health and disease.In this context, the Kuramoto model, a classical model apt todescribe oscillators’ dynamics, has been extended to capture thespatial displacement and the communication conditions in suchbrain network. Starting from a previous work in this ?eld ,we analyze this modi?ed model and compare it with otherexisting large-scale models. In doing so, our aim is to promotea set of mathematical tools useful to better understand realexperimental data in neuroscience and estimate brain dynamics.

In this paper we analyze the performances of some popular algorithms used to solve distributed optimization problems involving quadratic cost functions in a multi agent system. Namely, we study the performances of standard consensus, accelerated consensus and ADMM.  We analyze the scalar quadratic function case, under different scenarios and with structured graphs. We find that accelerated consensus is the algorithm with the best performance in all the cases analyzed. On the other hand, ADMM has performance comparable to the accelerated consensus when the graph is scarcely connected, while for dense graphs its performance deteriorates and becomes worse than the one of standard consensus. The results therefore suggest that the choice of the algorithm to solve the problem we analyze strongly depends on the graph, and that accelerated consensus should always be preferred.

In recent years complex networks have gained in-
creasing attention in different fields of science and engineering.
The problem of controlling these networks is an interesting and
challenging problem to investigate. In this paper we look at the
controllability problem focusing on the energy needed for the
control. Precisely not only we want to analyze whether a network
can be controlled, but we also want to establish whether the
control can be performed using a limited amount of energy.
We restrict our study to irreducible and (marginally) stable
networks and we find that the leading right and left eigenvectors
of the network matrix play a crucial role in this analysis.
Interestingly, our results suggest the existence of a connection
between controllability and network centrality, a well-known
concept in network science. In case the network is reversible, the
latter connection involves the PageRank, an extensively studied
type of centrality measure. Finally, the proposed results are
applied to examples concerning random graphs.

Recently, several working implementations of inband full-duplex wireless systems have been presented, where the same node can transmit and receive simultaneously in the same frequency band. The introduction of such a possibility at the physical layer could lead to improved performance but also poses several challenges at the MAC layer. In this paper, an innovative mechanism of channel contention in full-duplex OFDM wireless networks is proposed. This strategy is able to ensure efficient transmission scheduling with the result of avoiding collisions and effectively exploiting full-duplex opportunities. As a consequence, considerable performance improvements are observed with respect to standard and state-of-the-art MAC protocols for wireless networks, as highlighted by extensive simulations performed in ad hoc wireless networks with varying number of nodes.

In many applications, and especially thosewhere batch processes are involved, a target scalar outputof interest is often dependent on one or more time seriesof data. With the exponential growth in data logging inmodern industries such time series are increasingly availablefor statistical modeling in soft sensing applications. In orderto exploit time series data for predictive modelling, it isnecessary to summarise the information they contain as aset of features to use as model regressors. Typically thisis done in an unsupervised fashion using simple techniquessuch as computing statistical moments, principalcomponents or wavelet decompositions, often leading tosignificant information loss and hence suboptimal predictivemodels. In this paper, a functional learning paradigm isexploited in a supervised fashion to derive continuous,smooth estimates of time series data (yielding aggregatedlocal information), while simultaneously estimating a continuousshape function yielding optimal predictions. Theproposed Supervised Aggregative Feature Extraction (SAFE)methodology can be extended to support nonlinear predictivemodels by embedding the functional learning framework ina Reproducing Kernel Hilbert Spaces setting. SAFE has anumber of attractive features including closed form solutionand the ability to explicitly incorporate first and secondorder derivative information. Using simulation studies and apractical semiconductor manufacturing case study we highlightthe strengths of the new methodology with respect tostandard unsupervised feature extraction approaches.

In this paper, statistical Quality of Service provisioning in next generation heterogeneous mobile cellular networks is investigated. To this aim, any active entity of the cellular network is regarded as a queuing system, whose statistical QoS requirements depend on the specific application. In this context, by quantifying the performance in terms of effective capacity, we introduce a lower bound for the system performance that facilitates an efficient analysis. We exploit this analytical framework to give insights about the possible improvement of the statistical QoS experienced by the users if the current heterogeneous cellular network architecture migrates from a Half Duplex to a Full Duplex mode of operation. Numerical results and analysis are provided, where the network is modeled as a Mate?rn point processes with a hard core distance. The results demonstrate the accuracy and computational efficiency of the proposed scheme, especially in large scale wireless systems.