Hardware/firmware implementation of a soft sensor using an improved version of a fuzzy identification algorithm

This paper presents the design and implementation of an embedded soft sensor, i. e., a generic and autonomous hardware module, which can be applied to many complex plants, wherein a certain variable cannot be directly measured. It is implemented based on a fuzzy identification algorithm called ""Limited Rules"", employed to model continuous nonlinear processes. The fuzzy model has a Takagi-Sugeno-Kang structure and the premise parameters are defined based on the Fuzzy C-Means (FCM) clustering algorithm. The firmware contains the soft sensor and it runs online, estimating the target variable from other available variables. Tests have been performed using a simulated pH neutralization plant. The results of the embedded soft sensor have been considered satisfactory. A complete embedded inferential control system is also presented, including a soft sensor and a PID controller. (c) 2007, ISA. Published by Elsevier Ltd. All rights reserved.

Adaptive multiscale principal components analysis for online monitoring of wastewater treatment

Fault detection and isolation (FDI) are important steps in the monitoring and supervision of industrial processes. Biological wastewater treatment (WWT) plants are difficult to model, and hence to monitor, because of the complexity of the biological reactions and because plant influent and disturbances are highly variable and/or unmeasured. Multivariate statistical models have been developed for a wide variety of situations over the past few decades, proving successful in many applications. In this paper we develop a new monitoring algorithm based on Principal Components Analysis (PCA). It can be seen equivalently as making Multiscale PCA (MSPCA) adaptive, or as a multiscale decomposition of adaptive PCA. Adaptive Multiscale PCA (AdMSPCA) exploits the changing multivariate relationships between variables at different time-scales. Adaptation of scale PCA models over time permits them to follow the evolution of the process, inputs or disturbances. Performance of AdMSPCA and adaptive PCA on a real WWT data set is compared and contrasted. The most significant difference observed was the ability of AdMSPCA to adapt to a much wider range of changes. This was mainly due to the flexibility afforded by allowing each scale model to adapt whenever it did not signal an abnormal event at that scale. Relative detection speeds were examined only summarily, but seemed to depend on the characteristics of the faults/disturbances. The results of the algorithms were similar for sudden changes, but AdMSPCA appeared more sensitive to slower changes.

Online intra-task device scheduling for hard real-time systems

A large part of power dissipation in a system is generated by I/O devices. Increasingly these devices provide power saving mechanisms, inter alia to enhance battery life. While I/O device scheduling has been studied in the past for realtime systems, the use of energy resources by these scheduling algorithms may be improved. These approaches are crafted considering a very large overhead of device transitions. Technology enhancements have allowed the hardware vendors to reduce the device transition overhead and energy consumption. We propose an intra-task device scheduling algorithm for real time systems that allows to shut-down devices while ensuring system schedulability. Our results show an energy gain of up to 90% when compared to the techniques proposed in the state-of-the-art.

A lightweight distributed super peer election algorithm for unstructured dynamic P2P systems

Dissertação apresentada na Faculdade de Ciências e Tecnologia da Universidade Nova de Lisboa para obtenção do grau de Mestre em Engenharia Electrotécnica e de Computadores

Algorithm engineering for expression dag based number types

Magdeburg, Univ., Fak. für Informatik, Diss., 2015

Semi-online neural-Q_leaming for real-time robot learning

Reinforcement learning (RL) is a very suitable technique for robot learning, as it can learn in unknown environments and in real-time computation. The main difficulties in adapting classic RL algorithms to robotic systems are the generalization problem and the correct observation of the Markovian state. This paper attempts to solve the generalization problem by proposing the semi-online neural-Q_learning algorithm (SONQL). The algorithm uses the classic Q_learning technique with two modifications. First, a neural network (NN) approximates the Q_function allowing the use of continuous states and actions. Second, a database of the most representative learning samples accelerates and stabilizes the convergence. The term semi-online is referred to the fact that the algorithm uses the current but also past learning samples. However, the algorithm is able to learn in real-time while the robot is interacting with the environment. The paper shows simulated results with the "mountain-car" benchmark and, also, real results with an underwater robot in a target following behavior

Use of Hy's law and a new composite algorithm to predict acute liver failure in patients with drug-induced liver injury.

BACKGROUND & AIMS Hy's Law, which states that hepatocellular drug-induced liver injury (DILI) with jaundice indicates a serious reaction, is used widely to determine risk for acute liver failure (ALF). We aimed to optimize the definition of Hy's Law and to develop a model for predicting ALF in patients with DILI. METHODS We collected data from 771 patients with DILI (805 episodes) from the Spanish DILI registry, from April 1994 through August 2012. We analyzed data collected at DILI recognition and at the time of peak levels of alanine aminotransferase (ALT) and total bilirubin (TBL). RESULTS Of the 771 patients with DILI, 32 developed ALF. Hepatocellular injury, female sex, high levels of TBL, and a high ratio of aspartate aminotransferase (AST):ALT were independent risk factors for ALF. We compared 3 ways to use Hy's Law to predict which patients would develop ALF; all included TBL greater than 2-fold the upper limit of normal (×ULN) and either ALT level greater than 3 × ULN, a ratio (R) value (ALT × ULN/alkaline phosphatase × ULN) of 5 or greater, or a new ratio (nR) value (ALT or AST, whichever produced the highest ×ULN/ alkaline phosphatase × ULN value) of 5 or greater. At recognition of DILI, the R- and nR-based models identified patients who developed ALF with 67% and 63% specificity, respectively, whereas use of only ALT level identified them with 44% specificity. However, the level of ALT and the nR model each identified patients who developed ALF with 90% sensitivity, whereas the R criteria identified them with 83% sensitivity. An equal number of patients who did and did not develop ALF had alkaline phosphatase levels greater than 2 × ULN. An algorithm based on AST level greater than 17.3 × ULN, TBL greater than 6.6 × ULN, and AST:ALT greater than 1.5 identified patients who developed ALF with 82% specificity and 80% sensitivity. CONCLUSIONS When applied at DILI recognition, the nR criteria for Hy's Law provides the best balance of sensitivity and specificity whereas our new composite algorithm provides additional specificity in predicting the ultimate development of ALF.

Kalman Filtering in Online Paper Quality Assessment

Controlling the quality variables (such as basis weight, moisture etc.) is a vital part of making top quality paper or board. In this thesis, an advanced data assimilation tool is applied to the quality control system (QCS) of a paper or board machine. The functionality of the QCS is based on quality observations that are measured with a traversing scanner making a zigzag path. The basic idea is the following: The measured quality variable has to be separated into its machine direction (MD) and cross direction (CD) variations due to the fact that the QCS works separately in MD and CD. Traditionally this is done simply by assuming one scan of the zigzag path to be the CD profile and its mean value to be one point of the MD trend. In this thesis, a more advanced method is introduced. The fundamental idea is to use the signals’ frequency components to represent the variation in both CD and MD. To be able to get to the frequency domain, the Fourier transform is utilized. The frequency domain, that is, the Fourier components are then used as a state vector in a Kalman filter. The Kalman filter is a widely used data assimilation tool to combine noisy observations with a model. The observations here refer to the quality measurements and the model to the Fourier frequency components. By implementing the two dimensional Fourier transform into the Kalman filter, we get an advanced tool for the separation of CD and MD components in total variation or, to be more general, for data assimilation. A piece of a paper roll is analyzed and this tool is applied to model the dataset. As a result, it is clear that the Kalman filter algorithm is able to reconstruct the main features of the dataset from a zigzag path. Although the results are made with a very short sample of paper roll, it seems that this method has great potential to be used later on as a part of the quality control system.

Online Learning of Non-stationary Sequences

We consider an online learning scenario in which the learner can make predictions on the basis of a fixed set of experts. The performance of each expert may change over time in a manner unknown to the learner. We formulate a class of universal learning algorithms for this problem by expressing them as simple Bayesian algorithms operating on models analogous to Hidden Markov Models (HMMs). We derive a new performance bound for such algorithms which is considerably simpler than existing bounds. The bound provides the basis for learning the rate at which the identity of the optimal expert switches over time. We find an analytic expression for the a priori resolution at which we need to learn the rate parameter. We extend our scalar switching-rate result to models of the switching-rate that are governed by a matrix of parameters, i.e. arbitrary homogeneous HMMs. We apply and examine our algorithm in the context of the problem of energy management in wireless networks. We analyze the new results in the framework of Information Theory.

Semi-online neural-Q_leaming for real-time robot learning

Reinforcement learning (RL) is a very suitable technique for robot learning, as it can learn in unknown environments and in real-time computation. The main difficulties in adapting classic RL algorithms to robotic systems are the generalization problem and the correct observation of the Markovian state. This paper attempts to solve the generalization problem by proposing the semi-online neural-Q_learning algorithm (SONQL). The algorithm uses the classic Q_learning technique with two modifications. First, a neural network (NN) approximates the Q_function allowing the use of continuous states and actions. Second, a database of the most representative learning samples accelerates and stabilizes the convergence. The term semi-online is referred to the fact that the algorithm uses the current but also past learning samples. However, the algorithm is able to learn in real-time while the robot is interacting with the environment. The paper shows simulated results with the "mountain-car" benchmark and, also, real results with an underwater robot in a target following behavior

Online modeling with tunable RBF network

In this paper, we propose a novel online modeling algorithm for nonlinear and nonstationary systems using a radial basis function (RBF) neural network with a fixed number of hidden nodes. Each of the RBF basis functions has a tunable center vector and an adjustable diagonal covariance matrix. A multi-innovation recursive least square (MRLS) algorithm is applied to update the weights of RBF online, while the modeling performance is monitored. When the modeling residual of the RBF network becomes large in spite of the weight adaptation, a node identified as insignificant is replaced with a new node, for which the tunable center vector and diagonal covariance matrix are optimized using the quantum particle swarm optimization (QPSO) algorithm. The major contribution is to combine the MRLS weight adaptation and QPSO node structure optimization in an innovative way so that it can track well the local characteristic in the nonstationary system with a very sparse model. Simulation results show that the proposed algorithm has significantly better performance than existing approaches.

Online bayesian inference in some time-frequency representations of non-stationary processes

The use of Bayesian inference in the inference of time-frequency representations has, thus far, been limited to offline analysis of signals, using a smoothing spline based model of the time-frequency plane. In this paper we introduce a new framework that allows the routine use of Bayesian inference for online estimation of the time-varying spectral density of a locally stationary Gaussian process. The core of our approach is the use of a likelihood inspired by a local Whittle approximation. This choice, along with the use of a recursive algorithm for non-parametric estimation of the local spectral density, permits the use of a particle filter for estimating the time-varying spectral density online. We provide demonstrations of the algorithm through tracking chirps and the analysis of musical data.

A constrained recursive least squares algorithm for adaptive combination of multiple models

In this paper, we develop a novel constrained recursive least squares algorithm for adaptively combining a set of given multiple models. With data available in an online fashion, the linear combination coefficients of submodels are adapted via the proposed algorithm.We propose to minimize the mean square error with a forgetting factor, and apply the sum to one constraint to the combination parameters. Moreover an l1-norm constraint to the combination parameters is also applied with the aim to achieve sparsity of multiple models so that only a subset of models may be selected into the final model. Then a weighted l2-norm is applied as an approximation to the l1-norm term. As such at each time step, a closed solution of the model combination parameters is available. The contribution of this paper is to derive the proposed constrained recursive least squares algorithm that is computational efficient by exploiting matrix theory. The effectiveness of the approach has been demonstrated using both simulated and real time series examples.

Generating unconstrained two-dimensional non-guillotine cutting patterns by a Recursive Partitioning Algorithm

In this study, a dynamic programming approach to deal with the unconstrained two-dimensional non-guillotine cutting problem is presented. The method extends the recently introduced recursive partitioning approach for the manufacturer's pallet loading problem. The approach involves two phases and uses bounds based on unconstrained two-staged and non-staged guillotine cutting. The method is able to find the optimal cutting pattern of a large number of pro blem instances of moderate sizes known in the literature and a counterexample for which the approach fails to find known optimal solutions was not found. For the instances that the required computer runtime is excessive, the approach is combined with simple heuristics to reduce its running time. Detailed numerical experiments show the reliability of the method. Journal of the Operational Research Society (2012) 63, 183-200. doi: 10.1057/jors.2011.6 Published online 17 August 2011

FORA - A Fuzzy Set Based Framework for Online Reputation Management

The Social Web offers increasingly simple ways to publish and disseminate personal or opinionated information, which can rapidly exhibit a disastrous influence on the online reputation of organizations. Based on social Web data, this study describes the building of an ontology based on fuzzy sets. At the end of a recurring harvesting of folksonomies by Web agents, the aggregated tags are purified, linked, and transformed to a so-called fuzzy grassroots ontology by means of a fuzzy clustering algorithm. This self-updating ontology is used for online reputation analysis, a crucial task of reputation management, with the goal to follow the online conversation going on around an organization to discover and monitor its reputation. In addition, an application of the Fuzzy Online Reputation Analysis (FORA) framework, lesson learned, and potential extensions are discussed in this article.