418 research outputs found
Analysis and Development of Consensus-based Estimation Schemes
No full textIn the last few decades we assisted to an extraordinary expansion of the Internet and of wireless technologies. These interconnection technologies allow a continuously increasing number of devices to exchange information. This fact, together with the parallel increase in the availability of inexpensive nodes carrying a wide range of sensing capabilities, attracts the interest in developing
large-scale sensing platforms, which could be used to measure a variety of physical phenomena.
However, these huge networks of simple devices are subject to tight energy and bandwidth constraints, making efficient distributed estimation and data fusion algorithms a strong need, to avoid unmanageable computational and communicational burden on network bottleneck nodes.
%In this thesis we address some issues in this research avenue
In this thesis we address some issues in this field, presenting and analyzing distributed algorithms to solve specific distributed estimation problems and carrying out the analysis of some other recently-proposed algorithms.
To perform data fusion in a distributed fashion, we relay on consensus algorithms, namely algorithms that achieve agreement on a common value in the network.
Using consensus as a basic brick to build estimation algorithms we can take advantage of the solid understanding on this problem that many recent contributions deepened and sharpened, and we can leverage for our analysis on powerful and effective tools.
In the thesis we propose a distributed algorithm for offset removal and an algorithm for least-square identification of the wireless-channel parameters, motivated by the application of localization and tracking of a moving object. We present moreover a novel linear algebra inequality, useful in the analysis of randomized algorithms. This result comes into play when we carry out an analysis of a recently-proposed distributed Kalman filtering algorithm. Finally, we look at the intriguing set up of a network cooperation to estimate different but correlated quantities, proposing and analyzing a distributed algorithm that performs inference over a simple Gauss-Markov random field.Gli ultimi decenni sono stati segnati dallo straordinario sviluppo di Internet e dalla pervasiva diffusione della tecnologia wireless, consentendo ad un numero sempre maggiore di dispositivi di scambiare tra loro informazioni.
Questo fatto, assieme alla crescente disponibilità, a prezzi modici, di nodi equipaggiati con un'ampia varietà di dispositivi di misura, rende tecnologicamente concretizzabile l'idea di sviluppare grandi piattaforme di sensing, incaricate di monitorare qualsivoglia grandezza fisica.
Tuttavia, queste grandi reti di dispositivi estremamente semplici hanno stringenti vincoli sul consumo energetico e sulla banda di comunicazione, che rendono criticamente necessario lo sviluppo di tecniche efficienti per la stima e la data-fusion, così da evitare carichi computazionali e di comunicazione insostenibili ai colli di bottiglia della rete.
Questa tesi si propone di contribuire proprio in questo settore, presentando alcuni algoritmi per la soluzione distribuita di specifici problemi di stima ed analizzando le prestazioni di algoritmi recentemente proposti.
Strumento chiave nella decentralizzazione della stima è la teoria del consensus, che propone algoritmi in grado di portare l'intera rete a concordare su una specifica quantità. L'utilizzo di algoritmi di consensus come elemento base nella costruzione di algoritmi di stima ci consente di sfruttare la solida comprensione di questo problema, affinata dai molti risultati recentemente proposti in letteratura, e di sfruttare degli strumenti di analisi ben consolidati.
Nella tesi, motivati dal problema della localizzazione e del tracking di un oggetto, proponiamo un algoritmo per la compensazione degli offset ed un algoritmo per la stima ai minimi quadrati dei parametri caratterizzanti il canale wireless. Inoltre presentiamo un nuovo risultato di algebra lineare, utile nell'analisi di algoritmi randomizzati. Questo risultato giocherà un ruolo centrale nell'analisi qui proposta di un algoritmo distribuito per la stima alla Kalman. Infine, consideriamo l'interessante caso di una rete di sensori incaricata di stimare quantità diverse ma tra loro correlate e proponiamo un algoritmo per l'inferenza di un semplice campo di Gauss-Markov
A majorization inequality and its application to distributed Kalman filtering
No full textIn the analysis of a recently proposed distributed estimation algorithm based on the Kalman filtering and on gossip iterations, we needed to apply a new inequality which is valid for i.i.d. matrix valued random processes. This inequality can be useful in the analysis of the convergence rate of general jump Markov linear systems.
In this paper, we present this inequality. This is based on the theory of majorization and on its use in the analysis of the singular values. Finally we will show the impact of this inequality on the performance analysis of gossip based distributed Kalman filters
Comparing Different Individualized Black-Box Models for Insulin Pump Faults Detection on Artificial Pancreas Data
No full textObjective:
Prompt and accurate detection of insulin pump faults could be key in preventing sustained hyperglycemia and possibly ketoacidosis. Model-based fault detection techniques relay on patient models to warn the patient of a possible malfunctioning. Here, we want to compare four individualized models to understand if there is a preferable choice in terms of fault detection ability.
Method:
Individualized, linear, black-box parametric models (ARX, ARMAX, ARIMAX, BJ) are identified with BIC-based optimal order on 7 days of fault-free closed-loop data for 100 virtual subjects, using one of the most recent versions of the UVA-Padova T1 Diabetic Patient Simulator, that accounts for intra/inter-patient variability. An online prediction of up to PH=3h of glucose concentration, along with its confidence interval, is calculated through a Kalman filter based on the individualized model, running on data of past infused insulin (possibly affected by an unknown 6h insulin suppression), ingested meals and CGM values for 10 days. The real time fault detection algorithm raises an alarm if the predicted residual portion stays out its confidence interval for more than 15min.
Result:
The performance of the detection method is evaluated in terms of false positives per day and recall (FP/days, RE%), also taking into consideration the detection time. The performance is (0.12, 70%), (0.13, 85%), (0.17, 80%), (0.16, 78%), while the detection time is 236min, 242min, 237min, 238min for ARX, ARMAX, ARIMAX and BJ, respectively. If we compute the Euclidian distance from the optimal point (0, 100%), this metric will suggest ARMAX as the best model.
Conclusion:
Although ARMAX appears to be the best choice, the use of the other models only slightly impacts the fault detection performance and detection time
sj-docx-1-dst-10.1177_19322968221093665 – Supplemental material for Combined Use of Glucose-Specific Model Identification and Alarm Strategy Based on Prediction-Funnel to Improve Online Forecasting of Hypoglycemic Events
No full textSupplemental material, sj-docx-1-dst-10.1177_19322968221093665 for Combined Use of Glucose-Specific Model Identification and Alarm Strategy Based on Prediction-Funnel to Improve Online Forecasting of Hypoglycemic Events by Simone Faccioli, Francesco Prendin, Andrea Facchinetti, Giovanni Sparacino and Simone Del Favero in Journal of Diabetes Science and Technology</p
A Glucose-Specific Metric to Assess Predictors and Identify Models
No full textIn diabetes, the mean square error (MSE) metric is extensively used for assessing glucose prediction methods and identifying glucose models. One limitation of this metric is that, by equally treating errors in hypo-, eu-, and hyperglycemia, it is not able to weight the different clinical impact of errors in these three situations. In this paper, we propose a new cost function, which overcomes this limitation and can be used in place of MSE for several scopes, in particular for assessing the quality of glucose predictors and identifying glucose models. The new metric called glucose-specific MSE (gMSE) modifies MSE with a Clark error grid inspired penalty function, which penalizes overestimation in hypoglycemia and underestimation in hyperglycemia, i.e., the most harmful conditions on a clinical perspective. From a mathematical point of view, gMSE retains sensitivity of MSE and inherits some of its important mathematical features, in particular it has no local minima, simplifying the optimization. This makes it suitable for model identification purposes also. First, the goodness of it is demonstrated by means of three experiments, designed ad hoc to evidence its sensitivity to accuracy, precision, and distortion in glucose predictions. Second, a prediction assessment problem is presented, in which two real prediction profiles are compared. Results show that the MSE chooses the worst clinical situation, while gMSE correctly selects the situation with less clinical risk. Finally, we also demonstrate that models identified minimizing gMSE are more accurate in potentially harmful situations (hypo- and hyperglycemia) than those obtained by MS
Detection of Glucose Sensor Faults in an Artificial Pancreas via Whiteness Test on Kalman Filter Residuals
No full textContinuous Glucose Monitoring (CGM) sensors are key components in an artificial pancreas, an emerging tool for type 1 diabetes treatment. Malfunctioning of this component might reduce the efficacy of glucose control achieved by the system and even pose the safety of the patient at risk. Therefore, accurate and prompt detection of these anomalies is an important problem. This paper investigates a model-based method to detect CGM failures. Based on an individualized linear model of the subject, identified on hystorical data, the method predicts future glucose concentration through a one-step ahead Kalman predictor. The correct functioning of the system is then monitored using two different criteria: the first checks the magnitude of prediction residuals. The second checks the whiteness of the residuals through a correlogram test. The effectiveness of the two criteria is investigated and compared by performing tests on an in-silico dataset obtained by means of UVA/Padova Type 1 Diabetes simulator, accepted by the US Food and Drug Administration as a substitute of animal testing prior to artificial pancreas clinical trials on humans
dst-19-0142_appendix_revised – Supplemental material for Detection of Insulin Pump Malfunctioning to Improve Safety in Artificial Pancreas Using Unsupervised Algorithms
No full textSupplemental material, dst-19-0142_appendix_revised for Detection of Insulin Pump Malfunctioning to Improve Safety in Artificial Pancreas Using Unsupervised Algorithms by Lorenzo Meneghetti, Gian Antonio Susto and Simone Del Favero in Journal of Diabetes Science and Technology</p
Binder1 – Supplemental material for Detection of Insulin Pump Malfunctioning to Improve Safety in Artificial Pancreas Using Unsupervised Algorithms
No full textSupplemental material, Binder1 for Detection of Insulin Pump Malfunctioning to Improve Safety in Artificial Pancreas Using Unsupervised Algorithms by Lorenzo Meneghetti, Gian Antonio Susto and Simone Del Favero in Journal of Diabetes Science and Technology</p
Bayesian learning of probability density functions: a Markov chain Monte Carlo approach
No full textThe paper considers the problem of reconstructing a probability density function from a finite set of samples independently drawn from it.We cast the problem in a Bayesian setting where the unknown density is modeled via a nonlinear transformation of a Bayesian prior placed on a Reproducing Kernel Hilbert Space. The learning of the unknown density function is then formulated as a minimum variance estimation problem. Since this requires the solution of analytically intractable integrals, we solve this problem by proposing a novel algorithm based on the Markov chain Monte Carlo framework. Simulations are used to corroborate the goodness of the new approach.</p
- …
