This paper presents an automatic ECG-based emotion recognition algorithm for human emotion recognition. First, we adopt a musical induction method to induce participants' real emotional states and collect their ECG signals without any deliberate laboratory setting. Afterward, we develop an automatic ECG-based emotion recognition algorithm to recognize human emotions elicited by listening to music. Physiological ECG features extracted from the time-, and frequency-domain, and nonlinear analyses of ECG signals are used to find emotion-relevant features and to correlate them with emotional states. Subsequently, we develop a sequential forward floating selection-kernel-based class separability-based (SFFS-KBCS-based) feature selection algorithm and utilize the generalized discriminant analysis (GDA) to effectively select significant ECG features associated with emotions and to reduce the dimensions of the selected features, respectively. Positive/negative valence, high/low arousal, and four types of emotions (joy, tension, sadness, and peacefulness) are recognized using least squares support vector machine (LS-SVM) recognizers. The results show that the correct classification rates for positive/negative valence, high/low arousal, and four types of emotion classification tasks are 82.78, 72.91, and 61.52 percent, respectively.

Despite patients with Alzheimer's disease (AD) were reported of revealing gait disorders and balance problems, there is still lack of objective quantitative measurement of gait patterns and balance capability of AD patients. Based on an inertial-sensor-based wearable device, this paper develops gait and balance analyzing algorithms to obtain quantitative measurements and explores the essential indicators from the measurements for AD diagnosis. The gait analyzing algorithm is composed of stride detection followed by gait cycle decomposition so that gait parameters are developed from the decomposed gait details. On the other hand, the balance is measured by the sway speed in anterior-posterior (AP) and medial-lateral (ML) directions of the projection path of body's center of mass (COM). These devised gait and balance parameters were explored on twenty-one AD patients and fifty healthy controls (HCs). Special evaluation procedure including single-task and dual-task walking experiments for observing the cognitive function and attention is also devised for the comparison of AD and HC groups. Experimental results show that the wearable instrument with the designed gait and balance analyzing system is a promising tool for automatically analyzing gait information and balance ability, serving as assistant indicators for early diagnosis of AD.

This paper presents a wearable activity sensor system and a systematic activity classification scheme for the classification of human daily physical activities. The wearable activity sensor system, consisting of two activity sensor modules worn on users' dominant hand wrists and ankles, is used for collecting activity acceleration signals. The proposed activity classification scheme, including static/dynamic activity analysis, posture recognition, exercise classification, and ambulation classification, is capable of classifying time-series activity acceleration signals. The collected acceleration signals are classify into two categories by means of static/dynamic activity analysis. Posture recognition is applied for partitioning static signals into sitting and standing. Exercise classification and ambulation classification algorithms were used to classify dynamic activity signals. Our experimental results have successfully validated the effectiveness of the proposed wearable sensor system and the scheme of activity classification algorithms with an overall classification accuracy of 96% for seven types of daily activities.

This paper presents an integrated Bootstrap AdaBoost with k-nearest neighbor (KNN) algorithm for obstructive sleep apnea (OSA) screening based on electrocardiogram (ECG) recordings during sleep. The proposed method processes single-lead ECG recordings for predicting the presence of major sleep apnea and provides a minute-by-minute analysis of disordered breathing. In our analysis, 35 recordings collected from the Physionet Apnea-ECG database were used as the training/testing dataset. A variety of features based on RR interval, an ECG-derived respiratory signal, and cardiopulmonary coupling techniques were employed. A Bootstrap AdaBoost with k-dimensional tree KNN was used as the classifier, adopting feature selection to optimize classifier performance. The Bootstrap AdaBoost with KDKNN (BA-KDKNN) algorithm reached an accuracy of 91.95%, sensitivity of 99.36%, and specificity of up to 89.02% with ten features.

This paper presents an inertial-sensor-based wearable device and its associated stride detection algorithm to analyze gait information for patients with Alzheimer's disease (AD). The wearable gait analysis device is composed of a triaxial accelerometer, a microcontroller, and an RF wireless transmission module. To validate the effectiveness of the proposed device and algorithm, nine AD patients and three healthy controls were recruited to participate a gait analysis experiment. They were asked to mount the device on their foot and walk along a straight line of 40 meters at normal speed. The stride detection algorithm, consisting of procedures of data collection, signal preprocessing, and stride detection, has been developed for acquiring gait feature information from acceleration signals. The advantages of this wearable gait analysis device include the following: 1) It can be used anywhere without any external device, and 2) the stride detection algorithm can acquire gait feature information from acceleration signals automatically and effectively. Experimental results show that the AD patients exhibited a significantly shorter mean stride length and slower mean gait speed than those of the healthy controls. No significant differences in mean stride frequency and mean cadence were observed in the two groups. The variability in the percentage of the stance phase of the AD patients was slightly greater than that of the healthy controls. Based on the above results and discussions with physicians, we conclude that the proposed wearable gait analysis device is a promising tool for automatically analyzing gait information which can serve as indicators for early diagnosis of AD.

This paper presents a wearable module and neural-network-based activity classification algorithm for energy expenditure estimation. The purpose of our design is first to categorize physical activities with similar intensity levels, and then to construct energy expenditure regression (EER) models using neural networks in order to optimize the estimation performance. The classification of physical activities for EER model construction is based on the acceleration and ECG signal data collected by wearable sensor modules developed by our research lab. The proposed algorithm consists of procedures for data collection, data preprocessing, activity classification, feature selection, and construction of EER models using neural networks. In order to reduce the computational load and achieve satisfactory estimation performance, we employed sequential forward and backward search strategies for feature selection. Two representative neural networks, a radial basis function network (RBFN) and a generalized regression neural network (GRNN), were employed as EER models for performance comparisons. Our experimental results have successfully validated the effectiveness of our wearable sensor module and its neural-network-based activity classification algorithm for energy expenditure estimation. In addition, our results demonstrate the superior performance of GRNN as compared to RBFN.

This paper explores the detection of emotional levels, high, medium and low from ECG signals. Features of ECG are extracted from frequency domain, time domain and nonlinear method and normalized by z-score method. Then SFS feature selection is applied followed by LDA feature transformation. After that, the transformed features are applied to KNNR classifier. According to our results, it was found that subjects of different characteristics reveal different biosignal responses. While subjects are regarded as optimistic characteristics, they have higher responses on positive films. On the other hand, the pessimistic subjects have higher responses on negative films. When classifiers are established separately for optimistic and pessimistic subjects, we can achieve the recognition rate of 97.8%, where 7 features are selected, and 94.0%, where 8 features are selected, for optimistic and pessimistic groups respectively. When the classification is built from all the subjects, the recognition rate is reduced slightly, but it can still maintain a recognition rate of 90.4%.

This paper presents a digital circuit design approach for a commonly used activation function, hyperbolic tangent sigmoid functions, for neural networks. Our design concept for such a nonlinear function is to approximate the function of its first-order derivative by piece-wise linear functions first, then to obtain the estimate of the original function by integrating the approximated function of the first-order derivative by a digital circuit. The average error and maximum error of the proposed approximation approach are in the order of 10−3and 10−2, respectively in the software simulation. The hardware implementation of the proposed method consumes only one multiplication and one addition/subtraction ALU with the aid of resource sharing. The performance of our circuit has been validated by a neural network for a system identification problem in the software simulation.

This paper investigates the effectiveness of a single-channel speech enhancement system that contains spectral subtraction and voice activity detection algorithm for noise elimination. We first extract features from a noisy signal and use these features as the inputs of a recurrent neuro-fuzzy network for detecting the voice activities of the signal. Based on the detection, we describe the characteristics of the background noise of the speech segments by a minimum frequency energy (MFE) parameter and then apply spectral subtraction algorithms with the parameter to eliminate the noise. Our simulation results show that the proposed enhancement system with a nonlinear spectral subtraction algorithm has superior performance.

This paper presents a digital circuit design of a state-space recurrent neural network (RNN). The proposed digital circuit design separates the datapath of the state-space RNN into a linear subcircuit and a nonlinear subcircuit. The linear subcircuit is realized by a matrix-vector multiplier while the nonlinear subcircuit by a customized nonlinear function computing unit. The throughput rate of the proposed RNN circuit is 36060.5 times faster than that of the software simulation using MATLABreg. The proposed state-space RNN digital design methodology not only possesses the advantages including high computing speed, small area and portability, but also increases the possibility of using the digital RNN circuit in real-world dynamic problems.

This paper presents a Hammerstein-Wiener recurrent neural network with a parameter learning algorithm for identifying unknown dynamic nonlinear systems. The proposed recurrent neural network resembles the conventional Hammerstein-Wiener model that consists of a dynamic linear subsystem embedded between two static nonlinear subsystems. There are two novelties in our network: (1) the three subsystems are integrated into a single recurrent neural network whose output is the nonlinear transformation of a linear state-space equation; (2) the well-developed linear theory can be applied directly to the linear subsystem of the trained network to analyze its characteristics. In addition, we utilized the Stone-Weierstrass theorem to demonstrate the proposed network possesses the universal approximation capability. Finally, a computer simulation and comparisons with some existing models have been conducted to demonstrate the effectiveness of the proposed network and its parameter learning algorithm.

This paper presents a simplified support vector clustering (SVC) algorithm for improving the efficiency of the SVC training procedure. The cluster structure obtained by our proposed approach is controlled by two parameters: the parameter of kernel functions, denoted as q; and the percentage of data used to form the contour. The mechanisms we developed can efficiently search for suitable parameters without much trial-and-error effort for reaching a satisfactory clustering result. From observations of the behavior of the clustering, we found that 1) the search range of q is related to the densities of the clusters; 2) the number of boundary vectors has much relevance to the computation time; and 3) the shape of the original dataset affects the size of a reduced dataset. We have based our findings to develop a simplified SVC to identify optimal cluster configuration with suitable cluster contours. Computer simulations have been conducted on benchmark datasets to demonstrate the effectiveness of our proposed approach.

This paper presents a Hammerstein recurrent neurofuzzy network associated with an online minimal realization learning algorithm for dealing with nonlinear dynamic applications. We fuse the concept of states in linear systems into a neurofuzzy framework so that the whole structure can be expressed by a state-space representation. An online minimal realization learning algorithm has been developed to find a controllable and observable state-space model of minimal size from the input-output measurements of a given system. Such an idea can simultaneously resolve the problem of the determination of a minimal structure and the difficulty of network stability analysis. The advantages of our approach include: 1) our recurrent network is capable of translating the complicated dynamic behavior of a nonlinear system into a minimal set of linguistic fuzzy dynamical rules and into state-space representation as well and 2) an online minimal realization learning algorithm unifies an order determination algorithm, a hybrid parameter initialization method, and a recursive recurrent learning algorithm into a systematic procedure to identify a minimal structure with satisfactory performance. Performance evaluations on benchmark examples as well as real-world applications have successfully validated the effectiveness of our approach.

This paper focuses on the development of an effective cluster validity measure with outlier detection and cluster merging algorithms for support vector clustering (SVC). Since SVC is a kernel-based clustering approach, the parameter of kernel functions and the soft-margin constants in Lagrangian functions play a crucial role in the clustering results. The major contribution of this paper is that our proposed validity measure and algorithms are capable of identifying ideal parameters for SVC to reveal a suitable cluster configuration for a given data set. A validity measure, which is based on a ratio of cluster compactness to separation with outlier detection and a cluster-merging mechanism, has been developed to automatically determine ideal parameters for the kernel functions and soft-margin constants as well. With these parameters, the SVC algorithm is capable of identifying the optimal number of clusters with compact and smooth arbitrary-shaped cluster contours for the given data set and increasing robustness to outliers and noise. Several simulations, including artificial and benchmark data sets, have been conducted to demonstrate the effectiveness of the proposed cluster validity measure for the SVC algorithm.

This paper presents a Hammerstein neuro-fuzzy network with an online hybrid construction algorithm for dealing with dynamic applications. The proposed recurrent neuro-fuzzy system possesses two salient features: 1) it is capable of translating the complicated dynamic behavior of a system into a set of simple linguistic "dynamic" rules and into a state-space representation as well, and 2) with an automated hybrid construction algorithm, it can self-construct it network structure with a parsimonious size and satisfactory learning performance. Extensive computer simulations have been conducted to validate the effectiveness of the proposed approach for dynamic applications.

This paper presents a novel control design approach for controlling nonlinear dynamical systems using linear controllers with nonlinearity eliminators. Our control philosophy is to apply a divide-and-conquer strategy based on the assumption that an unknown system can be decomposed into two components: a static nonlinear model and a dynamic linear model. If the system modeling and the inverse of the nonlinear model are accurate, the compound model, the unknown system cascaded with the inverse model, will behave like the linear dynamic model. To effectively control such a linear model, well-developed linear control theory can directly be used in the feedback linear controller design. We first developed an ad hoc recurrent network structure that consists of a nonlinear model and a linear dynamic model. A fully automatic construction algorithm was devised to construct the recurrent structure. This algorithm integrates the order determination, parameter initialization and optimization, as well as the design procedure of the controller into a systematic framework. With this algorithm, the modeling and controller design procedures are totally exempted from trial-and-error approaches. Computer simulations on unknown system control problems have successfully validated the effectiveness of the proposed hybrid control scheme with superior control performance.

This paper presents a novel cluster validity method for the support vector clustering (SVC) algorithm to identify an optimal cluster configuration of a given data set. The SVC algorithm is a kernel-based clustering approach that groups a data set into clusters with irregular shapes. Without a priori knowledge of the data sets, a validity measure based on a ratio of cluster compactness to separation with outlier detection has been developed to automatically determine suitable parameters of the kernel functions and soft-margin constants as well. A novel validity measure has been developed to find optimal cluster configurations through an effective parameter searching algorithm. Computer simulations have been conducted on benchmark data sets to demonstrate the effectiveness of the proposed cluster validity method.

This paper presents a self-constructing recurrent neural network (SCRNN) capable of building itself with a compact structure from input-output measurements for identification of chaotic systems. The proposed SCRNN is constituted by a static nonlinear network cascaded with a linear dynamic network. A unified learning algorithm consisting of two mechanisms, a hybrid weight initialization method and a parameter optimization method, has been developed for the structure and parameter identification. With this learning algorithm, the SCRNN is exempted from trial and error in structure initialization as well as parameterization. Computer simulations on discrete-time chaotic systems, including logistic and Henon mappings, validate that the proposed SCRNN is capable of capturing the dynamical behavior of chaotic systems with a compact network size.

This paper presents a self-adaptive learning algorithm for dynamic system identification using a novel recurrent neural network with minimal representation. The proposed algorithm consists of two mechanisms, a minimal realization technique based on Markov parameters and a recursive parameter learning method on the ordered derivatives, for the minimal order identification and parameter optimization, respectively. Computer simulations on unknown dynamic system identification using the proposed approach have successfully validated: 1) the order of the recurrent network representation is minimal, and 2) the proposed network is able to closely capture the dynamical behavior of the unknown system with a satisfactory performance.

This paper presents a hybrid predictor for the CATS (competition on artificial time series) benchmark. Considering the time series as a sum of two components: the major trend and a residual series, we tackled the prediction problem by a hybrid predictor consisting of two models - a kernel regression model and a recurrent neuro-fuzzy model. The kernel regression model based on Gaussian function expansions was first applied to predict the major trend of the time series. The time series was sectioned into several data sets to obtain the best-fitting regression model. Subsequently, the recurrent neuro-fuzzy model associated with a learning algorithm was used to predict the dynamics of the residual series. The learning algorithm has been developed to construct a minimum size of the recurrent model in state-space representation. The best prediction results were presented and discussed.

This paper presents a self-regulating clustering algorithm (SRCA) that is capable of identifying the cluster configuration without a priori knowledge regarding the given data set. The proposed SRCA integrates growing, merging, and splitting mechanisms into a systematic framework to identify the minimal cluster configuration. A novel idea of cluster boundary estimation has been proposed to effectively perform the three mechanisms. A virtual cluster spread coupled with a regulating vector enables the proposed SRCA to reveal the compact cluster configuration which may close to the true one. Computer simulations have been conducted to demonstrate the effectiveness of the proposed SRCA in terms of a minimal error of cluster estimation.

This work presents a novel adaptive recurrent neuro-fuzzy filter (ARNFF) for speech enhancement in noisy environment. The speech enhancement scheme consists of two microphones that receive a primary and a reference input source respectively, and the proposed ARNFF that attenuates the noise corrupting the original speech signal in the primary channel. The ARNFF is a connectionist network that can be translated effortlessly into a set of dynamic fuzzy rules and state-space equations as well. An effective learning algorithm, consisting of a clustering algorithm for the structure learning and a recurrent learning algorithm for the parameter learning, is adopted from our previous research for the ARNNF construction. From our computer simulations and comparisons with some existing filters, the advantages of the proposed ARNFF for noisy speech enhancement include: 1) a more compact filter structure, 2) no a priori knowledge needed for the exact lagged order of the input variables, 3) a better performance in long-delay environment.

This paper presents a self-adaptive recurrent neuro-fuzzy inference system (R-SANFIS) for dealing with dynamic problems. The proposed recurrent system possesses two salient features: 1) it incorporates fuzzy basis functions (FBFs) with dynamic elements for better approximation of nonlinear dynamic functions, and 2) it is capable of translating the complicated behaviors of dynamic systems into a set of simple linguistic "dynamic" rules and state-space equations as well. A systematic self-adaptive learning algorithm has been developed to construct the R-SANFIS with a parsimonious network structure and fast parameter learning convergence. Computer simulations and the performance comparisons with some existing recurrent networks on identification and control of nonlinear dynamic systems have been conducted to validate the effectiveness of the proposed R-SANFIS.

This paper presents the utilization of a self-adaptive recurrent neuro-fuzzy control as a feedforward controller and a proportional-plus-derivative (PD) control as a feedback controller for controlling an autonomous underwater vehicle (AUV) in an unstructured environment. Without a priori knowledge, the recurrent neuro-fuzzy system is first trained to model the inverse dynamics of the AUV and then utilized as a feedforward controller to compute the nominal torque of the AUV along a desired trajectory. The PD feedback controller computes the error torque to minimize the system error along the desired trajectory. This error torque also provides an error signal for online updating the parameters in the recurrent neuro-fuzzy control to adapt in a changing environment. A systematic self-adaptive learning algorithm, consisting of a mapping-constrained agglomerative clustering algorithm for the structure learning and a recursive recurrent learning algorithm for the parameter learning, has been developed to construct the recurrent neuro-fuzzy system to model the inverse dynamics of an AUV with fast learning convergence. Computer simulations of the proposed recurrent neuro-fuzzy control scheme and its performance comparison with some existing controllers have been conducted to validate the effectiveness of the proposed approach.

This paper presents the utilization of a self-adaptive recurrent neuro-fuzzy control as a feedforward controller and a proportional-plus-derivative (PD) control as a feedback controller for controlling an autonomous underwater vehicle (AUV) in an unstructured environment. Without a priori knowledge, the recurrent neuro-fuzzy system is first trained to model the inverse dynamics of the AUV and then it utilized as a feedforward controller to compute the nominal torque of the AUV along a desired trajectory. The PD feedback controller computes the error torque to minimize the system error along the desired trajectory. This error torque also provides an error signal for online updating the parameters in the recurrent neuro fuzzy control to adapt in a changing environment. A systematic self-adaptive learning algorithm, consisting of a mapping-constrained agglomerative clustering algorithm for the structure learning and a recursive recurrent learning algorithm for the parameter learning, was developed to construct the recurrent neuro-fuzzy system to model the inverse dynamics of an AUV with fast learning convergence. Computer simulations of the proposed recurrent neuro-fuzzy control scheme and its performance comparison with an adaptive controller were conducted to validate the effectiveness of the proposed approach.

This paper presents a self-adaptive neuro-fuzzy inference system (SANFIS) that is capable of self-adapting and self-organizing its internal structure to acquire a parsimonious rule-base for interpreting the embedded knowledge of a system from the given training data set. A connectionist topology of fuzzy basis functions with their universal approximation capability is served as a fundamental SANFIS architecture that provides an elasticity to be extended to all existing fuzzy models whose consequent could be fuzzy term sets, fuzzy singletons, or functions of linear combination of input variables. Without a priori knowledge of the distribution of the training data set, a novel mapping-constrained agglomerative clustering algorithm is devised to reveal the true cluster configuration in a single pass for an initial SANFIS construction, estimating the location and variance of each cluster. Subsequently, a fast recursive linear/nonlinear least-squares algorithm is performed to further accelerate the learning convergence and improve the system performance. Good generalization capability, fast learning convergence and compact comprehensible knowledge representation summarize the strength of SANFIS. Computer simulations for the Iris, Wisconsin breast cancer, and wine classifications show that SANFIS achieves significant improvements in terms of learning convergence, higher accuracy in recognition, and a parsimonious architecture.

Examines several clustering methods for structure learning in constructing efficient neuro-fuzzy systems. The structure learning establishes the internal structure (i.e., the number of term sets and fuzzy-rule base generation) of a given neuro-fuzzy architecture. The fundamental ideas of existing rule generation algorithms are addressed and discussed. Performance of the neuro-fuzzy systems established from these clustering methods is validated through computer simulations of the classification problem of IRIS and the control example of an autonomous underwater vehicle.

Presents a systematic approach for developing a concise self-adaptive neuro-fuzzy inference system (SANFIS) with a fast hybrid parameter learning algorithm for online learning the control knowledge for autonomous underwater vehicles (AUV). The multi-layered structure of SANFIS incorporates fuzzy basis functions for better function approximations. Based on the need of different applications, we investigate three SANFIS structures with three different types of fuzzy IF-THEN-rule-based models and cast the rule formation problem as a clustering problem. A recursive least squares algorithm and a modified Levenberg-Marquardt algorithm with limited memory are exploited to accelerate the learning process. Thus, incorporating an online clustering technique, a fast hybrid learning procedure and rule examination, the SANFIS is capable of self-organizing and self-adapting its internal structure for learning the required control knowledge for an AUV to follow desired trajectories. Computer simulations for modeling a control system for an AUV have been conducted to validate the effectiveness of the proposed SANFIS.

This paper presents a systematic and fast learning algorithm for developing a parsimonious internal structure for self-adaptive neuro-fuzzy inference system (SANFIS). The rule extraction problem is cast as a clustering problem so that the number of rules and the number of term sets for input and output variables can be determined in an efficient and systematic way. The consequent of SANFIS could be fuzzy term sets, fuzzy singleton values, or functions of linear combination of input variables. Without a prior knowledge of the distribution of the training data set, the proposed mapping-constrained agglomerative clustering algorithm is able to reveal the true number of clusters and simultaneously estimate the centers and variances of the clusters for constructing an initial SANFIS structure in a single pass. Next, a fast linear/nonlinear parameter optimization algorithm is performed to further accelerate the learning convergence and improve the system performance.

Controlling autonomous underwater vehicles (AUVs) in an uncertain and unstructured environment presents many challenging control problems. Model-based control strategies have been used with limited success. The paper presents an online self-organizing neuro-fuzzy control that serves as a better alternative control scheme in controlling AUVs. The proposed self-organizing neuro-fuzzy controller is a six-layer feedforward neural network that is capable of self-constructing and self-restructuring its internal node connectivity and learning the parameters of each node based on incoming training data. Computer simulations have been conducted to validate the performance of the proposed neuro-fuzzy controller and an experimental verification has been scheduled to verify if on ODIN, an autonomous underwater vehicle developed at the University of Hawaii.

Presents an online self-adaptive neuro-fuzzy control that serves as a better alternative control scheme in controlling autonomous underwater vehicles (AUVs) in an uncertain and unstructured environment. The proposed self-adaptive neuro-fuzzy controller is a five-layer feedforward neural network that implements fuzzy basis function (FBF) expansions and is capable of self-constructing and self-restructuring its internal node connectivity and learning the parameters of each node based on incoming training data. Computer simulations have been conducted to validate the performance of the proposed neuro-fuzzy controller and an experimental verification has been scheduled to verify it on ODIN, an autonomous underwater vehicle developed at the University of Hawaii.