Distributed reconfigurable control strategies for switching topology networked multi-agent systems.

In this paper, distributed control reconfiguration strategies for directed switching topology networked multi-agent systems are developed and investigated. The proposed control strategies are invoked when the agents are subject to actuator faults and while the available fault detection and isolation (FDI) modules provide inaccurate and unreliable information on the estimation of faults severities. Our proposed strategies will ensure that the agents reach a consensus while an upper bound on the team performance index is ensured and satisfied. Three types of actuator faults are considered, namely: the loss of effectiveness fault, the outage fault, and the stuck fault. By utilizing quadratic and convex hull (composite) Lyapunov functions, two cooperative and distributed recovery strategies are designed and provided to select the gains of the proposed control laws such that the team objectives are guaranteed. Our proposed reconfigurable control laws are applied to a team of autonomous underwater vehicles (AUVs) under directed switching topologies and subject to simultaneous actuator faults. Simulation results demonstrate the effectiveness of our proposed distributed reconfiguration control laws in compensating for the effects of sudden actuator faults and subject to fault diagnosis module uncertainties and unreliabilities.

An ensemble of dynamic neural network identifiers for fault detection and isolation of gas turbine engines.

In this paper, a new approach for Fault Detection and Isolation (FDI) of gas turbine engines is proposed by developing an ensemble of dynamic neural network identifiers. For health monitoring of the gas turbine engine, its dynamics is first identified by constructing three separate or individual dynamic neural network architectures. Specifically, a dynamic multi-layer perceptron (MLP), a dynamic radial-basis function (RBF) neural network, and a dynamic support vector machine (SVM) are trained to individually identify and represent the gas turbine engine dynamics. Next, three ensemble-based techniques are developed to represent the gas turbine engine dynamics, namely, two heterogeneous ensemble models and one homogeneous ensemble model. It is first shown that all ensemble approaches do significantly improve the overall performance and accuracy of the developed system identification scheme when compared to each of the stand-alone solutions. The best selected stand-alone model (i.e., the dynamic RBF network) and the best selected ensemble architecture (i.e., the heterogeneous ensemble) in terms of their performances in achieving an accurate system identification are then selected for solving the FDI task. The required residual signals are generated by using both a single model-based solution and an ensemble-based solution under various gas turbine engine health conditions. Our extensive simulation studies demonstrate that the fault detection and isolation task achieved by using the residuals that are obtained from the dynamic ensemble scheme results in a significantly more accurate and reliable performance as illustrated through detailed quantitative confusion matrix analysis and comparative studies.

Hybrid fault diagnosis of nonlinear systems using neural parameter estimators.

This paper presents a novel integrated hybrid approach for fault diagnosis (FD) of nonlinear systems taking advantage of both the system's mathematical model and the adaptive nonlinear approximation capability of computational intelligence techniques. Unlike most FD techniques, the proposed solution simultaneously accomplishes fault detection, isolation, and identification (FDII) within a unified diagnostic module. At the core of this solution is a bank of adaptive neural parameter estimators (NPEs) associated with a set of single-parameter fault models. The NPEs continuously estimate unknown fault parameters (FPs) that are indicators of faults in the system. Two NPE structures, series-parallel and parallel, are developed with their exclusive set of desirable attributes. The parallel scheme is extremely robust to measurement noise and possesses a simpler, yet more solid, fault isolation logic. In contrast, the series-parallel scheme displays short FD delays and is robust to closed-loop system transients due to changes in control commands. Finally, a fault tolerant observer (FTO) is designed to extend the capability of the two NPEs that originally assumes full state measurements for systems that have only partial state measurements. The proposed FTO is a neural state estimator that can estimate unmeasured states even in the presence of faults. The estimated and the measured states then comprise the inputs to the two proposed FDII schemes. Simulation results for FDII of reaction wheels of a three-axis stabilized satellite in the presence of disturbances and noise demonstrate the effectiveness of the proposed FDII solutions under partial state measurements.

Gene interaction networks based on kernel correlation metrics.

In this paper, a Kernel correlation coefficient (KCC) method is proposed to elucidate the gene nonlinear relationships as a distance metric. To evaluate the performance of this nonlinear distance measure, a biological network of the Gaussian Kernel on a public dataset of yeast genes is constructed by using a graph theory. Specifically, the distribution and properties of this new measure are analysed and compared with the classical Pearson correlation method. The reliability and advantages of our proposed Kernel correlation metric is verified and shown formally on ten showcases of the DREAM (Dialogue for Reverse Engineering Assessments and Methods) project. Test experiment results demonstrate that the proposed Kernel correlation coefficient measure has a strong capability in identifying interaction genes, and that the proposed method can detect accurately the key genes and functional interactions (also known as the cliques) as compared to the commonly used Pearson correlation and Mutual Information measures.

A recurrent neural-network-based sensor and actuator fault detection and isolation for nonlinear systems with application to the satellite's attitude control subsystem.

This paper presents a robust fault detection and isolation (FDI) scheme for a general class of nonlinear systems using a neural-network-based observer strategy. Both actuator and sensor faults are considered. The nonlinear system considered is subject to both state and sensor uncertainties and disturbances. Two recurrent neural networks are employed to identify general unknown actuator and sensor faults, respectively. The neural network weights are updated according to a modified backpropagation scheme. Unlike many previous methods developed in the literature, our proposed FDI scheme does not rely on availability of full state measurements. The stability of the overall FDI scheme in presence of unknown sensor and actuator faults as well as plant and sensor noise and uncertainties is shown by using the Lyapunov's direct method. The stability analysis developed requires no restrictive assumptions on the system and/or the FDI algorithm. Magnetorquer-type actuators and magnetometer-type sensors that are commonly employed in the attitude control subsystem (ACS) of low-Earth orbit (LEO) satellites for attitude determination and control are considered in our case studies. The effectiveness and capabilities of our proposed fault diagnosis strategy are demonstrated and validated through extensive simulation studies.

Constructive feedforward neural networks using hermite polynomial activation functions.

In this paper, a constructive one-hidden-layer network is introduced where each hidden unit employs a polynomial function for its activation function that is different from other units. Specifically, both a structure level as well as a function level adaptation methodologies are utilized in constructing the network. The functional level adaptation scheme ensures that the "growing" or constructive network has different activation functions for each neuron such that the network may be able to capture the underlying input-output map more effectively. The activation functions considered consist of orthonormal Hermite polynomials. It is shown through extensive simulations that the proposed network yields improved performance when compared to networks having identical sigmoidal activation functions.

New training strategies for constructive neural networks with application to regression problems.

Regression problem is an important application area for neural networks (NNs). Among a large number of existing NN architectures, the feedforward NN (FNN) paradigm is one of the most widely used structures. Although one-hidden-layer feedforward neural networks (OHL-FNNs) have simple structures, they possess interesting representational and learning capabilities. In this paper, we are interested particularly in incremental constructive training of OHL-FNNs. In the proposed incremental constructive training schemes for an OHL-FNN, input-side training and output-side training may be separated in order to reduce the training time. A new technique is proposed to scale the error signal during the constructive learning process to improve the input-side training efficiency and to obtain better generalization performance. Two pruning methods for removing the input-side redundant connections have also been applied. Numerical simulations demonstrate the potential and advantages of the proposed strategies when compared to other existing techniques in the literature.

A neural-network appearance-based 3-D object recognition using independent component analysis.

This paper presents results on appearance-based three-dimensional (3-D) object recognition (3DOR) accomplished by utilizing a neural-network architecture developed based on independent component analysis (ICA). ICA has already been applied for face recognition in the literature with encouraging results. In this paper, we are exploring the possibility of utilizing the redundant information in the visual data to enhance the view based object recognition. The underlying premise here is that since ICA uses high-order statistics, it should in principle outperform principle component analysis (PCA), which does not utilize statistics higher than two, in the recognition task. Two databases of images captured by a CCD camera are used. It is demonstrated that ICA did perform better than PCA in one of the databases, but interestingly its performance was no better than PCA in the case of the second database. Thus, suggesting that the use of ICA may not necessarily always give better results than PCA, and that the application of ICA is highly data dependent. Various factors affecting the differences in the recognition performance using both methods are also discussed.

Application of adaptive constructive neural networks to image compression.

The objective of the paper is the application of an adaptive constructive one-hidden-layer feedforward neural networks (OHL-FNNs) to image compression. Comparisons with fixed structure neural networks are performed to demonstrate and illustrate the training and the generalization capabilities of the proposed adaptive constructive networks. The influence of quantization effects as well as comparison with the baseline JPEG scheme are also investigated. It has been demonstrated through several experiments that very promising results are obtained as compared to presently available techniques in the literature.

Experimental results on discrete-time nonlinear adaptive tracking control of a flexible-link manipulator.

The aim of this paper is to develop and implement a nonlinear adaptive control scheme for a single-link flexible manipulator. The controller is designed based on a discrete-time nonlinear model of the arm. The model is derived by using the forward difference method (Euler approximation). The output redefinition concept is then used so that the associated zero dynamics corresponding to the new output is guaranteed to be exponentially stable. An indirect adaptive linearizing controller is developed for the resulting minimum phase system where the "payload mass" is assumed to be unknown but its upper bound is assumed to be known a priori. The performance of the adaptively controlled closed-loop system is investigated by both numerical simulations and experimental results. The proposed controller is also compared experimentally with those of nonadaptive feedback linearization and conventional proportional derivative (PD) control strategies.

Identification of a two-link flexible manipulator using adaptive time delay neural networks.

This paper deals with identification of a two-link flexible manipulator belonging to a class of multi-input, multi-output (MIMO) nonlinear systems, by using adaptive time delay neural networks (ATDNNs). Two neuro-dynamic identifiers are proposed. The capabilities of the proposed structures for representing the nonlinear input-output map of the flexible manipulator are shown analytically. Selection criteria for specifying the fixed structural parameters as well as the adaptation laws for updating the adjustable parameters of the networks are provided. During identification, the two-link flexible manipulator is under nonlinear control and the input-output data sets are generated for different desired trajectories. Simulation results reveal that the proposed neuro-dynamic structures are capable of successfully identifying a highly nonlinear system without any a priori information about the nonlinearities of the system and without any off-line training.

Neural network based control schemes for flexible-link manipulators: simulations and experiments.

This paper presents simulation and experimental results on the performance of neural network-based controllers for tip position tracking of flexible-link manipulators. The controllers are designed by utilizing the modified output re-definition approach. The modified output re-definition approach requires only a priori knowledge about the linear model of the system and no a priori knowledge about the payload mass. Four different neural network schemes are proposed. The first two schemes are developed by using a modified version of the 'feedback-error-learning' approach to learn the inverse dynamics of the flexible manipulator. Both schemes require only a linear model of the system for defining the new outputs and for designing conventional PD-type controllers. This assumption is relaxed in the third and fourth schemes. In the third scheme, the controller is designed based on tracking the hub position while controlling the elastic deflection at the tip. In the fourth scheme which employs two neural networks, the first network (referred to as the 'output neural network') is responsible for specifying an appropriate output for ensuring minimum phase behavior of the system. The second neural network is responsible for implementing an inverse dynamics controller. The performance of the four proposed neural network controllers is illustrated by simulation results for a two-link planar flexible manipulator and by experimental results for a single flexible-link test-bed. The networks are all trained and employed as online controllers and no off-line training is required.

An Adaptive Structure Neural Networks with Application to EEG Automatic Seizure Detection.

This paper introduces a new algorithm for adaptively adjusting the structure of a multi-layer back-propagation network. The proposed algorithm belongs to the class of neuron generating strategies as opposed to the class of neuron pruning strategies. Initially a "small" multi-layer perceptron network is selected. The stabilized error is used as an index to determine whether the network needs to generate a new neuron or not. If after a period of learning the error is stabilized, but the error is larger than a desired value, then new neuron(s) is (are) generated. The new neurons are placed at locations that contribute most to the network error behavior through the fluctuation in their input weight vectors. Among the features of the new architecture are its improved performance and generalization capabilities compared to a standard fixed-structure back-propagation network. Application to an electroencephalogram (EEG) automatic epileptic seizure detection is presented to illustrate advantages and capabilities of the proposed algorithm. Using an actual data from five patients it is shown that the proposed approach correctly identifies all true seizures that are also identified by an expert physician. The new algorithm provides a reduction of 60-70% in the training epochs as compared to a back-propagation algorithm. Furthermore, it is shown that by utilizing a new training algorithm it is possible to reduce the false seizure detections to zero while resulting in a 5.1% error in identifying the true seizures. Copyright 1996 Elsevier Science Ltd