Real-time performance is essential for practical deployment of object detection on edge devices, where high processing speed and low latency are paramount. This paper introduces a novel approach aimed at boosting real-time object detection while strictly adhering to computational constraints. A structured state space sequence model, Mamba, is strategically embedded in the early stages of the backbone network to capture long-range dependencies, thereby enhancing the model’s representation capability. Given the limitations of Mamba in directional perception, a lightweight spatial attention mechanism is introduced to integrate global context into each spatial location. Additionally, a computationally efficient module inspired by the Ghost module is developed to reduce resource demands. This dual-strategy approach optimizes both performance and efficiency in real-time object detection. Extensive experiments demonstrate the superiority of this proposed approach; on the Microsoft Common Objects in Context (MS COCO) dataset, it achieves a +1.6 AP (Average Precision) improvement over state-of-the-art methods, reaching 41.1 AP with minimal added model complexity on the nano scale. The effectiveness and efficiency of each component are further substantiated through ablation studies on the Pascal Visual Object Classes (Pascal VOC dataset). To verify the universality of the proposed method, this study selects underwater object detection, characterized by an extremely complex background environment, as the other validation scenario. Through the application of this proposed approach to underwater object detection, a state-of-the-art result of 69.5 AP was obtained on the Detecting Underwater Objects (DUO) dataset, exceeding that of You Only Look Once Detector version 11 (YOLO11) by +0.3 AP. Code: https://github.com/chenjie04/Hybrid-YOLO .

This article proposes a fuzzy optimal fault-tolerant control scheme with prescribed performance for three-dimensional trajectory tracking control of underactuated autonomous underwater vehicles (AUVs) in complex ocean environments, especially in the presence of unknown actuator faults and dynamic uncertainties. The scheme has the following features. 1) The underactuation problem is overcome by defining new system outputs. 2) An error transformation method that removes restrictions on initial error conditions is adopted to achieve global prescribed performance control. 3) A fuzzy adaptive disturbance observer based on a generalized fuzzy hyperbolic model (GFHM) is proposed to estimate compound disturbances consisting of actuator faults, dynamic uncertainty parts and disturbances. 4) Based on adaptive dynamic programming techniques, a single critic structure fuzzy optimal controller is developed using a GFHM to obtain an approximate optimal control law. In particular, actuator faults, dynamic uncertainties, and disturbances are uniformly handled by structuring an improved cost function and solving optimal control of the nominal system. Through analysis, the proposed scheme can ensure that all signals of the AUV closed-loop system are uniformly ultimately bounded and the original tracking errors can remain within prescribed boundaries. Pure software and hardware-in-the-loop simulations demonstrate the effectiveness and advantages of the proposed scheme.

Optical imagery ship detection has achieved significant developments recently. However, accurate detection in complex scenes and for different-scale ships remains a vital challenge. To solve the above issues, in this article, we propose the asymmetric aggregation feature pyramid network (A2FPN), incorporating top-down semantic aggregation and bottom-up detail enhancement to propagate semantic and detailed information across different feature levels. In particular, the higher-level hierarchical features propagate global semantic information to the lower-level hierarchical features, successively enhancing the discriminative ability of each level of hierarchical features. After that, the lower-level hierarchical features with abundant semantic information are also aggregated successively to the higher-level hierarchical features through the augmentation path, enriching the details of each level of hierarchical features. Considering the real-time requirements of ship detection, we replace the original path aggregation feature pyramid network (FPN) of YOLOX with the proposed A2FPN and develop a ship detection model termed asymmetric aggregation network (A2Net). Extensive experiments are performed on the three commonly used ship detection datasets, ShipRSImageNet, Seaships7000, and HRSC2016. Quantitative and qualitative results demonstrate that A2Net outperforms the state-of-the-art methods.

Stereo vision, a popular depth estimation technology in computing vision, finds wide-ranging applications in embedded systems, including robotics vision and autonomous driving. These applications demand both high accuracy and fast processing speeds. To address hardware limitations, most current embedded systems rely on nonlearning algorithms for fast matching, sacrificing accuracy. Some recent studies have explored using convolutional neural networks (CNNs) to improve matching accuracy, but the computational load of existing learning-based systems hampers real-world applicability. This article presents significant contributions: 1) a novel stereo matching framework that greatly enhances accuracy on real-time embedded platforms and 2) a two-pronged approach combining a nonlearning-based algorithm and a lightweight super-resolution residual neural network (sRRNet). The nonlearning-based algorithm yields a low-resolution disparity map, while the lightweight sRRNet generates a high-resolution disparity map. Experimental results on benchmark data demonstrate that the proposed method achieves a low matching error rate of 5.17% and a real-time processing speed of 51 fps using the embedded Jetson AGX GPU. The proposed method outperforms all existing real-time embedded systems.

AUV play an important role in the exploration and utilization of the ocean. Underactuated autonomous underwater vehicles (AUVs) are widely used in these missions due to the associated low manufacturing costs, low power consumption, and high reliability. Among many control problems of underactuated AUVs, the research of trajectory tracking and formation control methods has become a hot topic in recent years. In order to obtain superior control performance, research on trajectory tracking and formation control of underactuated AUVs based on intelligent control methods should be further advanced and deepened. In this context, we present a survey of intelligent trajectory tracking and formation control of underactuated AUVs, which is beneficial to researchers in this field. To facilitate a comprehensive understanding of the subject matter, we review some preliminary knowledge in trajectory tracking and formation control of underactuated AUVs. We also highlight research problems and challenges, including external disturbances, system characteristics, and system faults. We review recent advances in intelligent trajectory tracking and formation control of underactuated AUVs and analyze and discuss their characteristics. Moreover, we put forward some prospects of advanced control techniques, especially artificial intelligence-based techniques for underactuated AUVs.

We are pleased to announce the publication of the Special Issue on Hybrid Control of Autonomous Mobile Robots: Architectures, Algorithms and Applications. The control problem of Autonomous Mobile Robots (AMR) in a dynamic environment is a fundament problem that has been receiving much attention from researchers from the world. The main issue here is how to obtain accurate, flexible, and reliable navigation? To perform a navigation task efficiently and effectively, the robot must have perception, decision-making and action capacities for interacting with the environment. The type and complexity of control architecture are usually related to the complexity of the environment and the task at hand. Navigation methods are classified into two main categories, namely, global planning methods (deliberative navigation) and local planning methods (reactive navigation). The main advantage of local planning methods is that they do not require a priori knowledge on the environment model and sometimes without the explicit model of the robot. In the recent years, several local planning methods have been developed. Most of them are based on artificial potential field, fuzzy logic and artificial neural networks. These methods are generally applicable to unknown environments and can be easily adapted to dynamically changing environments. However, such methods frequently suffer from the problem of local. In addition, the actual trajectory is not optimal in terms of distance and/or travel-time due to lack of global vision on the environment. In global planning methods, a navigation task can be achieved in two phases, namely, trajectory planning and tracking phases. Trajectory planning of a robot revolves around fulfilling some performance criteria (distance, travel-time, and energy consumption) and satisfying a certain number of constraints (geometric, kinematic, and/or dynamic). This ensures a safe and fast navigation solution taking into consideration kinematic and dynamic capacities of the robot, and the constraints related to the environment. However, these methods do not adapt to the dynamic of the environment (unexpected obstacles) or completely unknown environment. As regards to the trajectory planning, several approaches whereby the trajectory is generally made up of line segments connected via tangential circular arcs have been proposed. Most of these works deal with minimum-time trajectory-planning problems, under linear/angular velocity bounds of the platform. Some performance techniques have been developed to reach the goal as quickly as possible by smoothing transitions, thus achieving continuous-curvature trajectories. Concerning the problem of trajectory tracking, it revolves around following a reference trajectory by minimizing the position, orientation and sometimes speeds errors while maintaining the robot's stability. Many control methods have been proposed; some of them are the classic PID control, Lyapunov-based nonlinear control, sliding mode control, and fuzzy logic control. According to the available information on the navigation environment, methods of the first or second group are selected more often. This leads for three classes of control architectures, namely, reactive, deliberative and hybrid ones. Reactive control architectures are based on the “Sense & Act” principle that combines trajectory planning and its execution at the same level. Generally speaking, they are composed of a set of specific behavioral modules (task-specific behaviors). This allows the robot to make real-time decisions based on local perception and reactive interactions required in unknown and dynamically changing environments. The reference of most proposed solutions is the Subsumption Architecture which can be divided into two main classes based on competitive or cooperative mechanisms between behaviors modules. Deliberative control architecture is based on “Sense, Plan & Act” principle used in fully known environments. In fact, the robot model must be known and continually updated to plan the robot's actions. In this approach, one or more trajectories are first planned. Next, according to the actual state of the perceived information, the robot executes trajectory tracking strategies. Deliberative systems are considered as classical control architectures since they were the first to be tested. Given the drawbacks of the two types of methods, the combination of both types gives hybrid control architecture which enables navigation in partially known environments. This choice allows fast and reactive solution while avoiding unexpected obstacles and reducing the traveling time with introduction of partial knowledge of the environment. In fact, some interesting works adopting this approach have been reported in the literature. The last decade witnessed increasingly rapid progress in AI-powered hybrid control of AMR, mainly backed up by advances in the areas of artificial intelligence and deep learning. In particular, AI-based hybrid control architectures, convolutional, and recurrent neural networks, as well as the deep reinforcement learning paradigm have been proposed. These methodologies form a base for scene perception, path planning, behavior arbitration and motion control algorithms. Furthermore, modular perception-planning-action pipeline, where each module is built using deep learning methods which directly map sensory information to steering commands has been investigated. In this special issue, we included original contributions pertaining to architectures, algorithms and applications of hybrid control of AMR. We have performed a professional and strict review process in order to guarantee the quality of the special issue. We would also like to cordially thank all the reviewers who have participated in the review process of the articles submitted to this special issue, and the publishing team.

Exponential stability is often used to barricade stochastic disturbance in visual trajectory tracking of a robotic system. In this work, a nonlinear mathematical model to deal with the exponential stability of visual trajectory in the I-cub robot is developed. The main contributions of this article are: 1) exponential stability is studied for the first time in the literature for the neutral fractional stochastic differential equations (NFSDEs) driven by mixed Brownian motion (Bm) and subfractional Bm (sub-fBm); 2) existence and stability results are derived based on the Banach contraction principle, semi-group theory, and fractional calculus in stochastic settings; 3) stability results of mixed Bm and sub-fBm are established and applied to avoid dense environment stochastic disturbance in the visual trajectory of the I-cub robot; and 4) stability of sub-fBm is entrusted in the dense environment even in small particles from the shorter length. A weak solution of the proposed result ensures a sufficiently smooth solution of the considered robotic system. The obtained results are new and innovative in the sense that the proposed algorithms have several advantages for gaze shift framed to the I-cub robot, including helping to maintain the fixation of eye movement in the microscopic laboratory. Furthermore, the results are used to enhance the performance of decreasing rendering gaze shift to achieve high-quality visual tracking.

In this letter, a novel and effective algorithm based on Retinex model is proposed for low-light image enhancement, named Blind Adaptive Structure-Preserving Image Enhancement (BASSY). The low-light image enhancement is still a challenging task because the decomposition of images into light components and reflection components is an ill-posed problem. BASSY adopts a content-adaptive guided filtering based on local variances to estimate the proper illumination map. The salient features of the proposed approach are: (1) For the illumination component, the overall structure in the low-light image is preserved and the texture details are smoothed. (2) The reflectance is estimated without logarithmic transformation to reduce the computational burden and to avoid over-smoothing the reflectance component. (3) The adaptive gamma correction for the illumination map is used to reconstruct the enhanced image. (4) BASSY can be implemented efficiently due to the low computation complexity Ο(N). Experimental results on six public datasets show that the enhanced images by the BASSY exhibit higher naturalness and better visual quality than six state-of-the-art methods.

The fifth industrial revolution (Industry 5.0) integrates humans and machines to satisfy the increasing customization demands of the manufacturing complexity using an optimized robotized manufacturing process. Industry 5.0 make use of collaborative robots (cobots) for optimizing productivity and ensuring safety. At the same time, unmanned aerial vehicles (UAVs) are predicted to be the main part of industry 5.0 in the forthcoming days. Regardless of high mobility and energy-limited UAVs for wireless communication as significant advantages, different issues are also existing in the UAV networks, such as security, reliability, etc. Several research works have focused on resolving security issues in UAV communication to support safety-critical applications. With this motivation, this article presents an artificial intelligence-based UAV-borne secure communication with classification (AIUAV-SCC) framework for industry 5.0 environment. The proposed AIUAV-SCC model involves two major phases namely image steganography-based secure communication and deep learning (DL)-based classification. At the initial stage, a new image steganography technique with multilevel discrete wavelet transformation, quantum bacterial colony optimization based optimal pixel selection, and encryption processes take place. Next, in the second stage, the Bayesian optimization (BO)-based SqueezeNet model is applied for the classification of securely received UAV images where the parameters in the SqueezeNet method are optimally tuned by the utilize of the BO technique. To validate the performance of the presented model, extensive simulations are applied using the UC Merced dataset (UCM) aerial dataset and the outcomes are investigated under several dimensions. The outcomes make sure the goodness of the presented model on test UCM aerial dataset over the compared methods.

The output regulation approach has been extensively applied for the cooperative control of heterogeneous multiagent systems (MASs) with known dynamics, but rarely for MASs subject to stochastic and unknown dynamics. One challenging problem is that it is still unclear how to construct the regulator equations of stochastic MASs with unknown dynamics for dynamic exosystem compensation. Toward this end, this article develops a novel distributed control scheme for the approximate output regulation of discrete-time stochastic MASs subject to heterogeneous and unknown nonlinear dynamics. A distributed observer is designed for the exosystem-state estimation of agents, based upon which nonlinear regulator equations are constructed for the feedforward control design, thereby achieving distributed exosystem compensation. A high-order neural network is deployed to approximately solve the nonlinear regulator equations with unknown nonlinearity, and an adaptive control law is developed for the approximate output regulation of discrete-time stochastic MASs. Stability analysis shows that the closed-loop system is semiglobally uniformly ultimately bounded. Simulation results demonstrate the effectiveness and efficiency of the proposed control law.

This article reports our study on a reduced adaptive fuzzy decoupling control for our lower limb exoskeleton system which typically is a multi-input-multi-output (MIMO) uncertain nonlinear system. To show the applicability and generality of the proposed control methods, a more general MIMO uncertain nonlinear system model is considered. By decoupling control, the entire MIMO system is separated into several MISO subsystems. In our experiments, such a system may have problems (even unstable) if a traditional fuzzy approximator is used to estimate the complicated coupling terms. In this article, to overcome this problem, a reduced adaptive fuzzy system together with a compensation term is proposed. Compared to traditional approaches, the proposed fuzzy control approach can reduce possible chattering phenomena and achieve better control performance. By employing the proposed control scheme to an actual 2-DOF lower limb exoskeleton rehabilitation robot system, it can be seen from the experimental results that, as expected, it has good performance to track the model trajectory of a human walking gait. Therefore, it can be concluded that the developed approach is effective for the control of a lower limb exoskeleton system.

In this article, a global finite-time active disturbance rejection control (ADRC) scheme is proposed for tracking control of redundant parallel manipulators with unknown bounded uncertainties. This approach combines an ADRC and a global finite-time control for high accuracy trajectory tracking control. Based on the nonsingular fast terminal sliding mode control, the proposed approach can remove the condition in the original ADRC that the derivative of the uncertainties is required to be bounded. The extended state observer is employed to handle the real-time estimation of the total uncertainty. It can be found that the proposed scheme not only can converge fast to the semi-global finite-time stable equilibrium but also can have superior tracking control performance. In summary, compared to existing approaches, the proposed scheme can have several advantages, such as uncertainty rejection, easy implementation, robustness, chattering-free, high precision, and no need for prior knowledge of bounded uncertainties. The simulation results validate the effectiveness of the proposed method.

In this paper, the event-triggered consensus problem of linear multiagent systems with time-varying communication delays is addressed. Different from the existing event-triggered consensus results with communication delays, more general nonuniform time-varying communication delays are considered. To avoid the asynchronous phenomenon caused by nonuniform delays, a novel periodic switching controller is developed. Based on this controller, the resulting consensus error system can be modeled as a periodic switching system. Furthermore, the exponential stability of the consensus error system is derived by utilizing the Lyapunov approach and the dwell-time analysis method. Finally, an illustrative example is presented to demonstrate the effectiveness of the developed method.

This paper is concerned with the event-triggered dynamic output feedback control problem for Takagi-Sugeno (T-S) fuzzy systems under asynchronous network communications. Compared with the existing event-triggered output feedback results for the T-S fuzzy systems, two event-triggered mechanisms are predefined independently to check in an asynchronous manner whether the measurement output and the control input should be transmitted over networks or not. Consequently, network resources can be further saved. By introducing an auxiliary function in modeling, a delay system model is constructed. Then, a new stability criterion is presented such that the resulting closed-loop system is asymptotically stable. Furthermore, event-triggered parameters and controller gains can be codesigned if the related linear matrix inequalities are feasible. Finally, the validity of the theoretical result is illustrated by two examples.

The leader-following consensus problem for linear multi-agent systems with matched unknown nonlinear and actuator faults under switching topology is addressed in this paper. The main contributions are as follows: (1) In contrast to the existing results under switching topology, the unknown nonlinear considered in this paper are completely unknown; (2) The developed controller is capable of compensating for the actuator faults and the nonlinear simultaneously. To be more specific, by approximating the nonlinear by a fuzzy logical system (FLS) and by introducing switching mechanism in the distributed controller and adaptive update laws, a new FLS-based distributed adaptive controller is developed. By virtue of estimating the norm of weight vector in the FLS, the developed controller can compensate for unknown nonlinear under the actuator faults. In addition, it is proven that the developed controller can guarantee that the consensus errors are uniformly ultimately bounded. An illustrative example demonstrates the effectiveness and efficiency of the proposed method.

In this paper, the bearing-only formation control problem of multi-agent systems is addressed. The main contributions of the paper are twofold: (1) The local maximal clique graph instead of the global infinitesimal bearing rigidity graph is used to describe the formation configuration; (2) The proposed formation control law is globally stable. To be more specific, by considering the inter-agent communication topology satisfies the maximal clique graph condition, a cost function is designed based on the target formation. Next, the rotation bias of the final formation and the target formation are analyzed. Based on the negative gradient of the cost function, a globally stable distributed controller that only depends on the inter-bearing measurements is proposed and global convergence result and analysis are given. An illustrative example demonstrates the effectiveness and efficiency of the proposed control law.

In this paper, an universal adaptive fuzzy control (UAFC) scheme using output tracking error is proposed for practical tracking control of a class of nonlinear systems with unmeasured states and completely unknown perturbed dynamics including unknown dynamics and/or external disturbances. A tracking error system with measurable output is first derived from the output tracking problem, and unmeasured states are observed by an universal fuzzy state observer (UFSO), whereby adaptive fuzzy approximators and an universal adaptive gain are employed to estimate unknown dynamics and dominant unknown residuals, respectively. In conjunction with the rescaled UFSO and observation errors, the UAFC using output tracking error feedback is explicitly constructed, in a recursive manner, by employing the command filtered backstepping technique, whereby intermediate virtual signals and their first derivatives associated with complex dynamics can be reconstructed by second-order filters. Furthermore, adaptive mechanisms for fuzzy approximators and the universal gain pertaining to the UFSO and UAFC are derived from the Lyapunov synthesis. Theoretical analysis proves that all signals of the closed-loop system are bounded and the output tracking error and observation error can converge to an arbitrarily small region determined by a prescribed accuracy. Simulation results demonstrate the effectiveness and superiority of the proposed UAFC scheme.

This paper investigates the output consensus problem of heterogeneous stochastic nonlinear multiagent systems with directed communication topologies, with a view of making the outputs of a group of follower agents track the output of a leader. Fuzzy logic systems are applied to approximate the unknown nonlinear functions of agents. A special case that all followers can get access to the leader is first considered, and a novel decentralized adaptive fuzzy control law based on the output regulation framework is presented. Next, the proposed control scheme is further applied to design the distributed adaptive fuzzy control law for a more general case that only part of agents can get access to the leader. By applying Lyapunov stability analysis, it is shown that the outputs of followers will achieve consensus to a sufficient small bound of the output of the leader under the proposed control law. Finally, simulation results demonstrate that the proposed control law is effective and efficient. The developed distributed control scheme can be widely applied to solve the cooperative control problem of practical autonomous systems with uncertain dynamics such as synchronization of mechanical systems with vibration, formation control of autonomous underwater vehicles, etc.

Motivated by the challenging difficulty in tracking an uncertain marine vehicle (MV) with unknown dynamics and disturbances to any unmeasurable/unknown trajectory, which is unresolved, an adaptive universe-based fuzzy control (AUFC) scheme with retractable fuzzy partitioning (RFP) in global universe of discourse (UoD) is created to achieve global asymptotic model-free trajectory-independent tracking. By defining an error surface and intensively exploring the MV structure, tracking error dynamics are sufficiently trimmed via separating external unknowns including trajectory dynamics and disturbances from internal nonlinearities dependent on tracking errors. An innovative retractable fuzzy approximator (RFA) using the RFP is developed to estimate internal nonlinearities and does not require a priori knowledge on the UoD, thereby contributing to a globally adaptive approximation based control approach in conjunction with Lyapunov synthesis. Together with RFA residuals, external unknowns are globally dominated by adaptive universal compensators driven by tracking error surface. Eventually, tracking errors and their derivatives globally asymptotically converge to the origin and all other signals of the closed-loop system are bounded. Simulation studies demonstrate superior performance of the proposed AUFC scheme in terms of both tracking and approximation.

In this paper, for a general class of uncertain nonlinear (cascade) systems, including unknown dynamics, which are not feedback linearizable and cannot be solved by existing approaches, an innovative adaptive approximation-based regulation control (AARC) scheme is developed. Within the framework of adding a power integrator (API), by deriving adaptive laws for output weights and prediction error compensation pertaining to single-hidden-layer feedforward network (SLFN) from the Lyapunov synthesis, a series of SLFN-based approximators are explicitly constructed to exactly dominate completely unknown dynamics. By the virtue of significant advancements on the API technique, an adaptive API methodology is eventually established in combination with SLFN-based adaptive approximators, and it contributes to a recursive mechanism for the AARC scheme. As a consequence, the output regulation error can asymptotically converge to the origin, and all other signals of the closed-loop system are uniformly ultimately bounded. Simulation studies and comprehensive comparisons with backstepping- and API-based approaches demonstrate that the proposed AARC scheme achieves remarkable performance and superiority in dealing with unknown dynamics.

One key issue in content‐based image retrieval is to extract effective features so as to represent the visual content of an image. In this study, a new non‐negative sparse feature learning approach to produce a holistic image representation based on low‐level local features is presented. Specifically, a modified spectral clustering method is introduced to learn a non‐negative visual dictionary from local features of training images. A non‐negative sparse feature encoding method termed non‐negative locality‐constrained linear coding (NNLLC) is proposed to improve the popular locality‐constrained linear coding method so as to obtain more meaningful and interpretable sparse codes for feature representation. Moreover, a new feature pooling strategy named kMaxSum pooling is proposed to alleviate the information loss of the sum pooling or max pooling strategy, which produces a more effective holistic image representation and can be viewed as a generalisation of the sum and max pooling strategies. The retrieval results carried out on two public image databases demonstrate the effectiveness of the proposed approach.

Existing extreme learning algorithm have not taken into account four issues: 1) complexity; 2) uncertainty; 3) concept drift; and 4) high dimensionality. A novel incremental type-2 meta-cognitive extreme learning machine (ELM) called evolving type-2 ELM (eT2ELM) is proposed to cope with the four issues in this paper. The eT2ELM presents three main pillars of human meta-cognition: 1) what-to-learn; 2) how-to-learn; and 3) when-to-learn. The what-to-learn component selects important training samples for model updates by virtue of the online certainty-based active learning method, which renders eT2ELM as a semi-supervised classifier. The how-to-learn element develops a synergy between extreme learning theory and the evolving concept, whereby the hidden nodes can be generated and pruned automatically from data streams with no tuning of hidden nodes. The when-to-learn constituent makes use of the standard sample reserved strategy. A generalized interval type-2 fuzzy neural network is also put forward as a cognitive component, in which a hidden node is built upon the interval type-2 multivariate Gaussian function while exploiting a subset of Chebyshev series in the output node. The efficacy of the proposed eT2ELM is numerically validated in 12 data streams containing various concept drifts. The numerical results are confirmed by thorough statistical tests, where the eT2ELM demonstrates the most encouraging numerical results in delivering reliable prediction, while sustaining low complexity.

Multidocument summarization has gained popularity in many real world applications because vital information can be extracted within a short time. Extractive summarization aims to generate a summary of a document or a set of documents by ranking sentences and the ranking results rely heavily on the quality of sentence features. However, almost all previous algorithms require hand-crafted features for sentence representation. In this paper, we leverage on word embedding to represent sentences so as to avoid the intensive labor in feature engineering. An enhanced convolutional neural networks (CNNs) termed multiview CNNs is successfully developed to obtain the features of sentences and rank sentences jointly. Multiview learning is incorporated into the model to greatly enhance the learning capability of original CNN. We evaluate the generic summarization performance of our proposed method on five Document Understanding Conference datasets. The proposed system outperforms the state-of-the-art approaches and the improvement is statistically significant shown by paired t -test.

The age of online data stream and dynamic environments results in the increasing demand of advanced machine learning techniques to deal with concept drifts in large data streams. Evolving fuzzy systems (EFS) are one of recent initiatives from the fuzzy system community to resolve the issue. Existing EFSs are not robust against data uncertainty, temporal system dynamics, and the absence of system order, because a vast majority of EFSs are designed in the type-1 feedforward network architecture. This paper aims to solve the issue of data uncertainty, temporal behavior, and the absence of system order by developing a novel evolving recurrent fuzzy neural network, called evolving type-2 recurrent fuzzy neural network (eT2RFNN). eT2RFNN is constructed in a new recurrent network architecture, featuring double recurrent layers. The new recurrent network architecture evolves a generalized interval type-2 fuzzy rule, where the rule premise is built upon the interval type-2 multivariate Gaussian function, whereas the rule consequent is crafted by the nonlinear wavelet function. The eT2RFNN adopts a holistic concept of evolving systems, where the fuzzy rule can be automatically generated, pruned, merged, and recalled in the single-pass learning mode. eT2RFNN is capable of coping with the problem of high dimensionality because it is equipped with online feature selection technology. The efficacy of eT2RFNN was experimentally validated using artificial and real-world data streams and compared with prominent learning algorithms. eT2RFNN produced more reliable predictive accuracy, while retaining lower complexity than its counterparts.

In this paper the formation control problem of heterogeneous multi-agent systems is investigated. The formation control problem is first transformed to an output regulation problem, and then a distributed adaptive control law based on state feedback is designed to solve the problem. The salient feature of the developed control law is that the feedback gains are independent of the Laplacian matrix of the underlying system topology, which is of global nature. Furthermore, it is shown that all agents can form a formation and keep a desired relative position to a leader under a necessary and sufficient condition, and all feedback gains will approach some constant as time goes to infinity. An example demonstrates that the proposed control law is highly effective and efficient.

In recent years, Unmanned Air Vehicles (UAV) are gaining popularity among hobbyist, university-level researchers and military agencies. This paper presents the enhancement of a commercial-off-the-shelf (COTS) product that is capable of achieving semi-autonomous flight. The details of the selection of various components and software design are discussed. Take-off, hovering and landing controls for indoor flying capabilities are achieved by deriving the mathematical model of the quadcopter dynamics. Altitude control algorithm is implemented to ease the burden of the pilots. The final design of the quadcopter fulfils the flight capability and the required stability.

In this paper, the problem of asynchronous state estimate is investigated for a class of slowly switching system with time-varying overlapped detection delay. The switching signal satisfies average dwell time switching. The newly proposed overlapped detection delay allows a detection delay to begin before the end of one or more previous detection delays. The term asynchronous refers to the situation that the implemented estimator can be unmatched with the active subsystem (a subsystem is also called a mode of the switched system). A maximum total running time of the unmatched estimator is provided to ensure stability of the considered estimate error system. With such a limitation, asynchronous estimator is designed such that the final estimate error system is globally uniformly exponentially stable for certain average dwell time switching signal. A numerical example is provided to illustrate the effectiveness of the designed estimator and how the selected parameters affect the allowable total running time of the unmatched estimator.

Extractive summarization aims to generate a summary by ranking sentences, whose performance relies heavily on the quality of sentence features. In this paper, a document summarization framework based on convolutional neural networks is successfully developed to learn sentence features and perform sentence ranking jointly. We adapt the original CNN model to address a regression process for sentence ranking. Pre-trained word vectors are used to enhance the performance of our model. We evaluate our proposed method on the DUC 2002 and 2004 datasets covering single and multi-document summarization tasks respectively. The proposed system achieves competitive or even better performance compared with state-of-the-art document summarization systems.

Evolving intelligent system (EIS) is a machine learning algorithm, specifically designed to deal with learning from large data streams. Although the EIS research topic has attracted various contributions over the past decade, the issue of uncertainty, temporal system dynamic, and system order are relatively unexplored by existing studies. A novel EIS, namely evolving type-2 recurrent fuzzy neural network (eT2RFNN) is proposed in this paper. eT2RFNN features a novel recurrent network architecture, possessing double local recurrent connections. It generates a generalized interval type-2 fuzzy rule, where an interval type-2 multivariate Gaussian function constructs the rule premise, and the rule consequent is crafted by the nonlinear wavelet function. eT2RFNN adopts an open structure, where it can start learning process from scratch with an empty rule base. Fuzzy rules can be automatically generated according to degree of nonlinearity data stream conveys. It can performs a rule base simplification procedure by pruning and merging inactive, outdated and overlapping rules. eT2RFNN can deal with the high dimensionality problem, where an online dimensionality reduction method is integrated in the training process. The efficacy of the eT2RFNN has been numerically validated using two real-world data streams, where it provides high predictive accuracy, while retaining low complexity.

In this paper, a novel extreme learning machine based online multi-label classifier for real-time data streams is proposed. Multi-label classification is one of the actively researched machine learning paradigm that has gained much attention in the recent years due to its rapidly increasing real world applications. In contrast to traditional binary and multi-class classification, multi-label classification involves association of each of the input samples with a set of target labels simultaneously. There are no real-time online neural network based multi-label classifier available in the literature. In this paper, we exploit the inherent nature of high speed exhibited by the extreme learning machines to develop a novel online real-time classifier for multi-label data streams. The developed classifier is experimented with datasets from different application domains for consistency, performance and speed. The experimental studies show that the proposed method outperforms the existing state-of-the-art techniques in terms of speed and accuracy and can classify multi-label data streams in real-time.

Sentiment classification has been a very hot topic in the field of natural language processing (NLP) and understanding in recent years. Recurrent neural networks (RNN) is a widely used tool to deal with the classification problem of variable-length sentences. The standard RNN can only access the preceding context of a sentence. In this paper, a new architecture termed Comprehensive Attention Recurrent Neural Networks (CA-RNN) which can store preceding, succeeding and local contexts of any position in a sequence is developed. The bidirectional recurrent neural networks (BRNN) is used to access the past and future information while a convolutional layer is employed to capture local information. The standard RNN is also replaced by two recently emerged RNN variants, namely long short-term memory (LSTM) and gated recurrent unit (GRU), to enhance the effectiveness of the new architecture. Another salient feature of the proposed model is that it can be trained end-to-end without any human intervention. It is very easy to be implemented. We conduct experiments on several sentiment-labeled datasets and analysis tasks. Experiment results demonstrate that capturing comprehensive contextual information can significantly enhance the classification accuracy compared with the standard recurrent models and the new models can achieve competitive performance compared with the state-of-the-art approaches.

In this paper, a progressive learning algorithm for multi-label classification to learn new labels while retaining the knowledge of previous labels is designed. New output neurons corresponding to new labels are added and the neural network connections and parameters are automatically restructured as if the label has been introduced from the beginning. This work is the first of the kind in multi-label classifier for class-incremental learning. It is useful for real-world applications such as robotics where streaming data are available and the number of labels is often unknown. Based on the Extreme Learning Machine framework, a novel universal classifier with plug and play capabilities for progressive multi-label classification is developed. Experimental results on various benchmark synthetic and real datasets validate the efficiency and effectiveness of our proposed algorithm.

Classification involves the learning of the mapping function that associates input samples to corresponding target label. There are two major categories of classification problems: Single-label classification and Multi-label classification. Traditional binary and multi-class classifications are sub-categories of single-label classification. Several classifiers are developed for binary, multi-class and multi-label classification problems, but there are no classifiers available in the literature capable of performing all three types of classification. In this paper, a novel online universal classifier capable of performing all the three types of classification is proposed. Being a high speed online classifier, the proposed technique can be applied to streaming data applications. The performance of the developed classifier is evaluated using datasets from binary, multi-class and multi-label problems. The results obtained are compared with state-of-the-art techniques from each of the classification types.

This paper is concerned with asynchronous H∞state-feedback stabilization problem for a class of switched linear systems with detection delays. Without lose of generality, the switched signal is considered to satisfy dwell time switching. Time-varying overlapped detection delay, which is allowed to be overlapped with others and is much more challenging than the detection delay of overlapped-free, is introduced. By establishing the stability of the considered autonomous system, a controller is designed such that the closed-loop nominal system is globally uniformly exponentially stable. H∞stabilization controller is further designed for the considered system with additive disturbance. A numerical example illustrates that the theoretical results are excellent.

It is well known that the architecture of the extreme learning machine (ELM) significantly affects its performance and how to determine a suitable set of hidden neurons is recognized as a key issue to some extent. The leave-one-out cross-validation (LOO-CV) is usually used to select a model with good generalization performance among potential candidates. The primary reason for using the LOO-CV is that it is unbiased and reliable as long as similar distribution exists in the training and testing data. However, the LOO-CV has rarely been implemented in practice because of its notorious slow execution speed. In this paper, an efficient LOO-CV formula and an efficient LOO-CV-based ELM (ELOO-ELM) algorithm are proposed. The proposed ELOO-ELM algorithm can achieve fast learning speed similar to the original ELM without compromising the reliability feature of the LOO-CV. Furthermore, minimal user intervention is required for the ELOO-ELM, thus it can be easily adopted by nonexperts and implemented in automation processes. Experimentation studies on benchmark datasets demonstrate that the proposed ELOO-ELM algorithm can achieve good generalization with limited user intervention while retaining the efficiency feature.

In this paper, a novel adaptive robust online constructive fuzzy control (AR-OCFC) scheme, employing an online constructive fuzzy approximator (OCFA), to deal with tracking surface vehicles with uncertainties and unknown disturbances is proposed. Significant contributions of this paper are as follows: 1) unlike previous self-organizing fuzzy neural networks, the OCFA employs decoupled distance measure to dynamically allocate discriminable and sparse fuzzy sets in each dimension and is able to parsimoniously self-construct high interpretable T-S fuzzy rules; 2) an OCFA-based dominant adaptive controller (DAC) is designed by employing the improved projection-based adaptive laws derived from the Lyapunov synthesis which can guarantee reasonable fuzzy partitions; 3) closed-loop system stability and robustness are ensured by stable cancelation and decoupled adaptive compensation, respectively, thereby contributing to an auxiliary robust controller (ARC); and 4) global asymptotic closed-loop system can be guaranteed by AR-OCFC consisting of DAC and ARC and all signals are bounded. Simulation studies and comprehensive comparisons with state-of-the-arts fixed- and dynamic-structure adaptive control schemes demonstrate superior performance of the AR-OCFC in terms of tracking and approximation accuracy.

In this paper, an extreme learning control (ELC) framework using the single-hidden-layer feedforward network (SLFN) with random hidden nodes for tracking an unmanned surface vehicle suffering from unknown dynamics and external disturbances is proposed. By combining tracking errors with derivatives, an error surface and transformed states are defined to encapsulate unknown dynamics and disturbances into a lumped vector field of transformed states. The lumped nonlinearity is further identified accurately by an extreme-learning-machine-based SLFN approximator which does not require a priori system knowledge nor tuning input weights. Only output weights of the SLFN need to be updated by adaptive projection-based laws derived from the Lyapunov approach. Moreover, an error compensator is incorporated to suppress approximation residuals, and thereby contributing to the robustness and global asymptotic stability of the closed-loop ELC system. Simulation studies and comprehensive comparisons demonstrate that the ELC framework achieves high accuracy in both tracking and approximation.

Adaptive fuzzy controllers (AFCs), i.e., fuzzy controllers using heuristic expert knowledge and equipped with adaption ability, are widely used in the industry in the present days as they are capable of dealing with ill-formulated nonlinear systems (where exact mathematical formulas are not known) or partially described nonlinear systems. AFCs have been widely deployed to control many nonlinear systems with different types of critical nonlinearities. Over the last two decades, many different approaches toward adaptive fuzzy control have been reported. In this paper, we recapitulate the latest research works related to AFCs and provide an up-to-date survey on the latest development made in adaptive fuzzy control theory and applications based on different types of fuzzy systems considered and nonlinearities associated with the nonlinear system to be controlled. We categorize the different approaches based on different types of nonlinearities and consider basic nonlinear systems, nonlinear systems with external disturbances, nonlinear systems with some special type of nonlinearities, nonaffine nonlinear systems, and strict-feedback nonlinear systems. It turns out that many existing methods are subject to various limitations, and none of them is capable of controlling a general nonlinear systems. Hence, it is still an open problem to design and develop a universal AFC for a general class of nonlinear systems.

A research approach of crack detection of rotating shafts based on acoustic emission (AE) signals and machine learning is proposed in this paper. The relationship between crack intensity and domain features are investigated, and the features which could well indicate the crack condition are selected for modelling and crack prediction. Multiple Linear Regression (MLR), Artificial Neural Networks (ANN) and Adaptive Neural-Fuzzy Inference System (ANFIS) methods are used to establish the predictive correlation models by using selected features. A case study is carried out to emulate online working conditions of rotating shafts by using 10 normal shafts with 0.8mm – 8mm crack intensities. It is proved that AE signals can be used for earlier crack intensity detection, for example 0.8mm – 2.4 mm cracks can be fully detected according to experimental results in this study. Different modelling methods are also compared and discussed. Results show that ANFIS is a good choice in terms of overall predictive accuracy for earlier crack detection and prediction.

In this work, we discuss how the Lewinian model of experiential learning theory can be modeled in the framework of fuzzy logic. Fuzzy inference mechanism has been used to model the Lewinian model. Each stage of the Lewinian model has been modeled by appropriate step of a fuzzy inference mechanism.

In this paper, the motion dynamics of a large tanker is modeled by the generalized ellipsoidal function-based fuzzy neural network (GEBF-FNN). The reference model of tanker motion dynamics in the form of nonlinear difference equations is established to generate training data samples for the GEBF-FNN algorithm which begins with no hidden neuron. In the sequel, fuzzy rules associated with the GEBF-FNN-based model can be online self-constructed by generation criteria and parameter estimation, and can dynamically capture essential motion dynamics of the large tanker with high prediction accuracy. Simulation studies and comprehensive comparisons are conducted on typical zig-zag maneuvers with moderate and extreme steering, and demonstrate that the GEBF-FNN-based model of tanker motion dynamics achieves superior performance in terms of both approximation and prediction.

In this paper, a novel evolving fuzzy-rule-based classifier, termed parsimonious classifier (pClass), is proposed. pClass can drive its learning engine from scratch with an empty rule base or initially trained fuzzy models. It adopts an open structure and plug and play concept where automatic knowledge building, rule-based simplification, knowledge recall mechanism, and soft feature reduction can be carried out on the fly with limited expert knowledge and without prior assumptions to underlying data distribution. In this paper, three state-of-the-art classifier architectures engaging multi-input-multi-output, multimodel, and round robin architectures are also critically analyzed. The efficacy of the pClass has been numerically validated by means of real-world and synthetic streaming data, possessing various concept drifts, noisy learning environments, and dynamic class attributes. In addition, comparative studies with prominent algorithms using comprehensive statistical tests have confirmed that the pClass delivers more superior performance in terms of classification rate, number of fuzzy rules, and number of rule-base parameters.

This paper deals with tanker steering control based on a novel multiple-input multiple-output generalized ellipsoidal-basis-function-based fuzzy neural network (GEBF-FNN) with online updating of system structure and parameters. The main contributions of this paper are as follows. 1) A GEBF-FNN-based nonlinear steering model incorporating the nonlinearity underlying tanker dynamics is proposed. 2) The static local controller (SLC), whose controller gains are locally fixed with the initial forward speed and the desired heading for individual steering commands, is implemented. 3) The dynamic local controller (DLC) is further realized by employing adaptive controller gains pertaining to time-varying forward speed and heading dynamics. 4) The GEBF-FNN-based steering controller is developed by identifying a nonlinear mapping from the heading error, acceleration and forward speed to dynamic controller gains, and thereby contributing to a model-free adaptive control scheme. Simulation results and comprehensive studies on benchmark problems demonstrate that the GEBF-FNN-based model can capture the essential tanker dynamics, and the proposed SLC, DLC, and GEBF-FNN-based schemes achieve superior performance in terms of heading regulation and forward speed loss. In comparison with the SLC and traditional fuzzy controllers, the DLC and GEBF-FNN-based controllers achieve higher accuracy of heading regulation with less rudder efforts and minimal forward speed losses.

In this paper, a novel approach for detecting multiscale vehicles with time-varying vehicle features based on a multiscale and-or graph (AOG) model is proposed. Our approach consists of two steps, i.e., construction of a multiscale AOG model and an inference process for vehicle detection. The multiscale model uses global features to describe low-scale vehicles and local features to represent high-scale vehicles. Meanwhile, multiple appearances, such as sketch, flatness, texture, and color, are used to represent the global and local features. By virtue of the use of both global and local features as well as multiple appearances, our model is more suitable for describing multiscale vehicles in complex urban traffic conditions. Based on this multiscale model, an inference process using local features (local process) is integrated with a process using global features (global process) to detect multiscale vehicles. To evaluate the performance of our proposed method, a validation experiment, a quantitative evaluation, and a contrasting experiment are conducted. The experimental results show that our proposed approach can efficiently detect multiscale vehicles. In addition, the results also demonstrate that our approach is able to handle partial vehicle occlusion and various vehicle shapes and has great potential for real-world applications.

In this paper, traditional single-hidden layer feedforward network (SLFN) is extended to novel generalized SLFN (GSLFN) by employing polynomial functions of inputs as output weights connecting randomly generated hidden units with corresponding output nodes. The significant contributions of this paper are as follows: 1) a primal GSLFN (P-GSLFN) is implemented using randomly generated hidden nodes and polynomial output weights whereby the regression matrix is augmented by full or partial input variables and only polynomial coefficients are to be estimated; 2) a simplified GSLFN (S-GSLFN) is realized by decomposing the polynomial output weights of the P-GSLFN into randomly generated polynomial nodes and tunable output weights; 3) both P- and S-GSLFN are able to achieve universal approximation if the output weights are tuned by ridge regression estimators; and 4) by virtue of the developed batch and online sequential ridge ELM (BR-ELM and OSR-ELM) learning algorithms, high performance of the proposed GSLFNs in terms of generalization and learning speed is guaranteed. Comprehensive simulation studies and comparisons with standard SLFNs are carried out on real-world regression benchmark data sets. Simulation results demonstrate that the innovative GSLFNs using BR-ELM and OSR-ELM are superior to standard SLFNs in terms of accuracy, training speed, and structure compactness.

A novel meta-cognitive-based scaffolding classifier, namely Generic-Classifier (gClass), is proposed in this paper to handle non-stationary classification problems in the single-pass learning mode. Meta-cognitive learning is a breakthrough in the machine learning where the learning process is not only directed to craft learning strategies to exacerbate the classification rates, i.e., how-to-leam aspect, but also is focused to accommodate the emotional reasoning and commonsense of human being in terms of what-to-leam and when-to-learn facets. The crux of gClass is to synergize the scaffolding learning concept, which constitutes a well-known tutoring theory in the psychological literatures, in the how-to-learn context of meta-cognitive learning, in order to boost the learner's performance in dealing with complex data. A comprehensive empirical studies in time-varying datasets is carried out, where gClass numerical results are benchmarked with other state-of-the-art classifiers. gClass is, generally speaking, capable of delivering the most encouraging numerical results where a trade-off between predictive accuracy and classifier's complexity can be achieved.

This paper documents the development of a quad-copter with indoor control scheme and navigation capability. A stabilized flying control system including traditional PID controller, Raspberry Pi onboard flight computer and electronic speed controller was developed to provide the basic platform for the quad-copter. PID tuning was utilized optimize the performance. In addition, RGB and depth cameras and other sensors were deployed to enable remote semi-autonomous control.

Extreme Learning Machine (ELM) is a special single-hidden-layer feedforward neural networks with very fast learning speed and has attracted significant research attentions in recent years. The salient feature of ELM is that the input parameters can be randomly generated instead of being exhaustively tuned, and thus saving a great deal of computational expenses. However, the architecture of ELM has a great impact on its generalization performance and is traditionally determined by a trial and error manner. Therefore selecting an appropriate ELM architecture becomes the crucial problem in the successful application of ELM. In this paper, we propose a Robust Incremental ELM (RI-ELM), a constructive method where the hidden nodes are added one by one. We consider RI-ELM as a robust algorithm, because the suitable architecture is selected based on the Leave-One-Out (LOO) Cross-Validation procedure, a nearly unbiased and reliable criterion, but with notorious slow implementation speed. To tackle this speed issue, we propose an efficient formula that can incrementally update the LOO error with every new hidden node recruited, thus RI-ELM can secure the speed advantage of ELM and achieve good and robust performance. Furthermore, RI-ELM requires nearly zero user intervention since the architecture is automatically determined.

In this paper, an Extreme Learning Machine (ELM) based technique for Multi-label classification problems is proposed and discussed. In multi-label classification, each of the input data samples belongs to one or more than one class labels. The traditional binary and multi-class classification problems are the subset of the multi-label problem with the number of labels corresponding to each sample limited to one. The proposed ELM based multi-label classification technique is evaluated with six different benchmark multi-label datasets from different domains such as multimedia, text and biology. A detailed comparison of the results is made by comparing the proposed method with the results from nine state of the arts techniques for five different evaluation metrics. The nine methods are chosen from different categories of multi-label methods. The comparative results shows that the proposed Extreme Learning Machine based multi-label classification technique is a better alternative than the existing state of the art methods for multi-label problems.

An observer based adaptive iterative learning control (AILC) is proposed for MIMO nonlinear systems with delayed output in this paper. Since the system state vector is unavailable for measurement, we apply the state tracking error observer to solve the problem of unmeasurable system state vector for the design of AILC. By using the state tracking error observer, a mixed time-domain and s-domain technique is first applied to derive an output observation error model. The output observation error model will become a decoupled MIMO linear systems whose input vector is the system uncertain vector and each diagonal element is a stable transfer function with relative degree one. Then, the output observation error model is further transformed by introducing an averaging filter matrix and some auxiliary signal vectors so that the AILC can be implemented without using differentiators. Based on the derived output observation error model, an MIMO filtered fuzzy neural network using delayed state estimation vector and state estimation vector as the input vector is applied to approximate the unknown system nonlinear function vector. Besides, a normalization signal is applied as a bounding function to design a robust learning component for compensation of the lumped uncertainties vector caused by function approximation error vector, state estimation error vector and delayed system output vector. Finally, a stabilization learning component is used to guarantee the boundedness of internal signals. Based on Lyapunov-like analysis, it is shown that all the adjustable parameters as well as internal signals remain bounded for all iterations. The norm of output tracking error vector will asymptotically converge to a tunable residual set whose size depends on some design parameters of averaging filter.

In this paper, a fast learning methodology for neuro-fuzzy inference system (NFIS) referred to as meta-cognitive fuzzy extreme learning machine (McFELM) is proposed. It is based on the original OS-Fuzzy-ELM algorithm incorporating principles of human meta-cognition to make the learning more effective. McFELM has two components: the cognitive component and the meta-cognitive component. The cognitive component is a fuzzy extreme learning machine which learn sequential data in a one-by-one mode or a chunk-by-chunk mode with fixed or varying chunk size, while the meta-cognitive component controls the learning process of the cognitive component using a self-regulating mechanism to decide what-to-learn, when-to-learn, and how-to-learn. Unlike the OS-Fuzzy-ELM algorithm which uses all arriving samples to update the output weight matrix, the proposed algorithm employs different strategies namely sample deletion, sample reserve and sample learning strategy to decide whether the data will be deleted directly, reserved for later use or used immediately. Instantaneous error is used to select the best learning strategy. The evaluation of McFELM is presented doing simulations on a nonlinear system identification problem and a set of benchmark regression problems from UCI machine leaning repository. The results show that the proposed McFELM produces better performance compared with existing algorithms.

Artificial Neural Networks (ANN) is a major machine learning technique inspired by biological neural networks. However, the process of its parameter tuning is usually tedious and time consuming, and thus it becomes a major bottleneck for it being efficiently applied and used by nonexperts. In this paper, a novel ANN algorithm, termed as Automatic Regularized Extreme Learning Machine (AR-ELM), based on a Regularized Extreme Learning Machine (RELM) using ridge regression is proposed. It is a true automatic ANN learning algorithm in the sense that it can automatically identify the appropriate essential system parameter according to the input data without the need of user intervention. Since this method is based on a relatively straightforward formula, it can achieve very fast learning speed. The simulation results shows that the proposed AR-ELM algorithm can achieve comparable results to tedious cross-validation tuned RELM. Furthermore, we also systematically investigate one of the biggest concerns of ELM, its randomness nature, caused by randomly generated parameters.

This paper proves that adaptive neural plus proportional-derivative (PD) control can lead to semiglobal asymptotic stabilization rather than uniform ultimate boundedness for a class of uncertain affine nonlinear systems. An integral Lyapunov function-based ideal control law is introduced to avoid the control singularity problem. A variable-gain PD control term without the knowledge of plant bounds is presented to semiglobally stabilize the closed-loop system. Based on a linearly parameterized raised-cosine radial basis function neural network, a key property of optimal approximation is exploited to facilitate stability analysis. It is proved that the closed-loop system achieves semiglobal asymptotic stability by the appropriate choice of control parameters. Compared with previous adaptive approximation-based semiglobal or asymptotic stabilization approaches, our approach not only significantly simplifies control design, but also relaxes constraint conditions on the plant. Two illustrative examples have been provided to verify the theoretical results.

Novel constructive and destructive parsimonious extreme learning machines (CP- and DP-ELM) are proposed in this paper. By virtue of the proposed ELMs, parsimonious structure and excellent generalization of multiinput-multioutput single hidden-layer feedforward networks (SLFNs) are obtained. The proposed ELMs are developed by innovative decomposition of the recursive orthogonal least squares procedure into sequential partial orthogonalization (SPO). The salient features of the proposed approaches are as follows: 1) Initial hidden nodes are randomly generated by the ELM methodology and recursively orthogonalized into an upper triangular matrix with dramatic reduction in matrix size; 2) the constructive SPO in the CP-ELM focuses on the partial matrix with the subcolumn of the selected regressor including nonzeros as the first column while the destructive SPO in the DP-ELM operates on the partial matrix including elements determined by the removed regressor; 3) termination criteria for CP- and DP-ELM are simplified by the additional residual error reduction method; and 4) the output weights of the SLFN need not be solved in the model selection procedure and is derived from the final upper triangular equation by backward substitution. Both single- and multi-output real-world regression data sets are used to verify the effectiveness and superiority of the CP- and DP-ELM in terms of parsimonious architecture and generalization accuracy. Innovative applications to nonlinear time-series modeling demonstrate superior identification results.

This paper certifies that standard adaptive fuzzy control (AFC) can guarantee asymptotic stabilization performance rather than uniformly ultimately boundedness (UUB) even in the presence of fuzzy approximation errors (FAEs). Under a direct AFC scheme, the resulting optimal FAE is shown to be bounded by the norm of the plant state vector multiplying a globally invertible and nondecreasing function, which provides a pivotal property for asymptotic stability analysis. Without any additional control compensation, the closed-loop system is proved to be partially and asymptotically stable in the sense that all involved signals are UUB and the plant state variables converge to zero. The resulting control law is certainly continuous since it only contains an adaptive fuzzy system. Compared with previous adaptive approximation-based asymptotic stabilization approaches, the proposed approach not only simplifies control design, but also relaxes constraint conditions on the controlled plant. A simulation example of inverted pendulum control is provided to verify the discovery of this study.

In this paper, an enhanced adaptive fuzzy control (AFC) strategy with guaranteed convergence of an optimal fuzzy approximation error (FAE) is presented for a class of uncertain nonlinear systems in the general Brunovsky form. Based on the fuzzy logic system (FLS) with variable universes of discourse, relaxed sufficient conditions that guarantee the optimal FAE being convergent are given, and the upper bound of the optimal FAE is obtained. The control singularity problem resulting from the unknown affine term is resolved by a novel fuzzy approximation equation, and the parameter adaptive law of the FLS is derived by the Lyapunov synthesis. By means of the optimal FAE bound result, it is proved that the closed-loop system achieves partially asymptotic stability under a certain selection of control parameters. The proposed approach retains all advantages of a previous similar approach under relaxed constraint conditions. Thus, it provides a more flexible solution to the AFC with optimal FAE convergence. Simulation studies have demonstrated high-precision tracking performance with smooth control input of the proposed approach.

A major challenge faced by machine learning community is the decision making problems under uncertainty. Reinforcement Learning (RL) techniques provide a powerful solution for it. An agent used by RL interacts with a dynamic environment and finds a policy through a reward function, without using target labels like Supervised Learning (SL). However, one fundamental assumption of existing RL algorithms is that reward function, the most succinct representation of the designer's intention, needs to be provided beforehand. In practice, the reward function can be very hard to specify and exhaustive to tune for large and complex problems, and this inspires the development of Inverse Reinforcement Learning (IRL), an extension of RL, which directly tackles this problem by learning the reward function through expert demonstrations. IRL introduces a new way of learning policies by deriving expert's intentions, in contrast to directly learning policies, which can be redundant and have poor generalization ability. In this paper, the original IRL algorithms and its close variants, as well as their recent advances are reviewed and compared.

Function approximators have been extensively used in Reinforcement Learning (RL) to deal with large or continuous space problems. However, batch learning Neural Networks (NN), one of the most common approximators, has been rarely applied to RL. In this paper, possible reasons for this are laid out and a solution is proposed. Specifically, a Receding Horizon Cache (RHC) structure is designed to collect training data for NN by dynamically archiving state-action pairs and actively updating their Q-values, which makes batch learning NN much easier to implement. Together with Extreme Learning Machine (ELM), a new RL with function approximation algorithm termed as RHC and ELM based RL (RHC-ELM-RL) is proposed. A mountain car task was carried out to test RHC-ELM-RL and compare its performance with other algorithms.

In this paper, we propose a novel face descriptor for face recognition, named Local Line Derivative Pattern (LLDP). High-order derivative images in two directions are obtained by convolving original images with Sobel Masks. A revised binary coding function is proposed and three standards on arranging the weights are also proposed. Based on the standards, the weights of a line neighborhood in two directions are arranged. The LLDP labels in two directions are calculated with the proposed binary coding function and weights. The labeled image is divided into blocks where spatial histograms are extracted separately and concatenated into an entire histogram as features for recognition. The experiments on the FERET and Extended Yale B show superior performances of the proposed LLDP compared to other existing methods based on the LBP. The results prove that the LLDP has good robustness against expression, illumination and aging variations.

Every doctor needs to learn how to diagnose accurately and reliably. Based on observations and knowledge, they have to diagnose illnesses and give individual treatment to each patient. Although there are numerous medical books, records and courses assisting doctors with their deduction, the medical knowledge outdates quickly and cannot replace one's own experience. To handle this challenge, this paper applies the fast and accurate online self-organizing scheme for parsimonious fuzzy neural networks (FAOS-PFNN) to biomedical diagnosis and prediction. Unlike other fuzzy neural networks, the FAOS-PFNN is a more practical method which does not require structure identification in advance and can achieve a more compact network structure. The effectiveness of the FAOS-PFNN has been tested on diagnosis of breast cancer and prediction of Parkinson's Disease respectively. Simulation studies demonstrate that the FAOS-PFNN algorithm can efficiently and accurately diagnose and therefore improve the computer assisted medical diagnosis.

We present an adaptive control scheme in a neural based model to improve the performance of articulatory based speech synthesis. The model generates the articulatory trajectories for English plosive-vowel patterns through regional target approximation in the articulatory domain and categorical perception in the auditory domain. The proposed method can effectively model the variations of static vowel sounds and adapt to changes and uncertainties in the dynamic plosive sounds. Simulation studies demonstrate that the proposed scheme is able to estimate and regulate the control parameters in the articulatory synthesizer to produce smooth and authentic acoustic phonetic output.

In metal cutting processes, tool condition monitoring (TCM) plays an important role in maintaining the quality of surface finishing. Monitoring of tool wear in order to prevent surface damage is one of the difficult tasks in the context of TCM. Through early detection, high quality surface finishing and near-zero loss for potential failures can be ensured. Real-time/online tool degradation detection by using machine learning is highly desired. The ability to predict the tool wear, which is related to the remaining useful life of a tool, will improve efficiency and optimize tool usage while ensuring the quality of the work piece produced. In this paper, examine two popular methods of machine learning, namely the Adaptive Network Fuzzy Inference System (ANFIS) and Support Vector Machine (SVM) are used to estimate the tool wear and correlation models for tool wear estimation using ANFIS and SVM are estimated. A case study for six sets of ball nose cutters in a high speed milling machining process of Inconel 718 is carried out. Comparative studies of the two methods are carried out and experimental results analysed and discussed. In turns out that the accuracy of the ANFIS is generally better than the SVM whereas SVM is much faster than ANFIS in terms of speed.

We describe a neural based articulatory phonetic inversion model to improve the recognition of the acoustically varying vowels and the syllable initial plosives. The model uses a set of continuous valued articulatory phonetic features (APFs) to explore the interactions between the motor control of articulators and the acoustic phonetic events. We demonstrate that the neural model gives more accurate and robust recognition performance on the TIMIT sentences. The model offers two salient properties: it allows asynchronous feature changes at phoneme boundaries, and it accounts for the dual aspects of human speech production and perception through a heuristic learning algorithm during APFs mapping.

In this paper, a novel efficient learning algorithm towards self-generating fuzzy neural network (SGFNN) is proposed based on ellipsoidal basis function (EBF) and is functionally equivalent to a Takagi-Sugeno-Kang (TSK) fuzzy system. The proposed algorithm is simple and efficient and is able to generate a fuzzy neural network with high accuracy and compact structure. The structure learning algorithm of the proposed SGFNN combines criteria of fuzzy-rule generation with a pruning technology. The Kalman filter (KF) algorithm is used to adjust the consequent parameters of the SGFNN. The SGFNN is employed in a wide range of applications ranging from function approximation and nonlinear system identification to chaotic time-series prediction problem and real-world fuel consumption prediction problem. Simulation results and comparative studies with other algorithms demonstrate that a more compact architecture with high performance can be obtained by the proposed algorithm. In particular, this paper presents an adaptive modeling and control scheme for drug delivery system based on the proposed SGFNN. Simulation study demonstrates the ability of the proposed approach for estimating the drug's effect and regulating blood pressure at a prescribed level.

This paper describes a recurrent neural network (RNN) based articulatory-phonetic inversion (API) model for improved speech recognition. And a specialized optimization algorithm is introduced to enable human-like heuristic learning in an efficient data-driven manner to capture the dynamic nature of English speech pronunciations. The API model demonstrates superior pronunciation modeling ability and robustness against noise contaminations in large-vocabulary speech recognition experiments. Using a simple rescoring formula, it improves the hidden Markov model (HMM) baseline speech recognizer with consistent error rates reduction of 5.30% and 10.14% for phoneme recognition tasks on clean and noisy speech respectively on the selected TIMIT datasets. And an error rate reduction of 3.35% is obtained for the SCRIBE-TIMIT word recognition tasks. The proposed system qualifies as a competitive candidate for profound pronunciation modeling with intrinsic salient features such as generality and portability.

With no a priori knowledge of plant boundary functions, a novel direct adaptive fuzzy controller (AFC) for a class of single-input single-output (SISO) uncertain affine nonlinear systems is developed in this paper. Based on the theory of fuzzy logic systems (FLSs) with variable universes of discourse (UDs), sufficient conditions that guarantee that the optimal fuzzy approximation error (FAE) is locally convergent are given. By the use of the output tracking error and its derivatives as input variables and by the selection of suitable adjusting parameters, a variable UD FLS with an optimal FAE local convergence is constructed, and its parameter adaptive law is derived by virtue of the Lyapunov stability theorem. Under the assumption that the optimal FAE is bounded, it is proved that the closed-loop system is asymptotically stable in the sense that all variables are uniformly ultimately bounded and that the tracking errors converge to zero. The proposed approach eliminates the influence of the FAE on the tracking errors by means of the inherent mechanism of the variable UD FLS. Thus, it has the potential to achieve high control performance without additional compensation under only a few fuzzy rules. Simulation studies demonstrate the superiority of the proposed AFC in terms of the settling time, tracking accuracy, smoothness of the control input, and robustness against external disturbances and parameter variations.

This paper presents an intelligent control approach for blood pressure system using self-generating fuzzy neural networks (SGFNN). The proposed SGFNN is simple and effective and is able to generate a fuzzy neural network to model unknown nonlinearities of complex blood pressure system. This paper investigates the use of fuzzy neural network technique for modeling and automatic control of mean arterial pressure (MAP) through the intravenous infusion of sodium nitroprusside (SNP). Simulation studies based on a sensitive model of MAP illustrate the ability of the proposed approach in modeling drug delivery system and control postsurgical blood pressure.

This paper shows that the integration of statistical and connectionist methods can greatly enhance human-computer interaction through speech. The research approach is inspired by recent advances in high performance automatic speech recognition (ASR) systems and neurocognitive researches of natural language understanding (NLU). And a modest hybrid computational model is proposed and implemented to achieve intelligent spoken language understanding (SLU) in an information retrieval system.

Localization is an important research problem in WSNs (wireless sensor networks) and many WSN localization algorithms have been proposed in the literature. For a single sensor node whose anchors' positions are known, the location estimation error lower bound can be computed by using the CRLB (Cramer-Rao Lower Bound). However, it is still unclear to the research community what the localization error lower bound is from a network point of view. In this paper, we derive a lower bound of the expected localization error for a network whose sensor nodes and anchors are randomly distributed according to a Poisson point process. We show that the lower bound of the expected localization error for a network is a function of the anchor density and the variance of anchor-to-sensor distance measurements.

An Enhanced Online Self-organizing Fuzzy Neural Network (EOS-FNN) is proposed in this paper. The proposed algorithm can improve computational efficiency while achieving comparable performance and accuracy compared to other methods. The proposed EOS-FNN starts with an empty rule set and automatically generates fuzzy rules according to the proposed criteria during the learning process. All the parameters of the fuzzy rules are updated by the Extended Kalman Filter (EKF) method. Nonlinear time-series prediction processes are used to evaluate the performance of the proposed EOS-FNN algorithm with a comparison to other popular algorithms including DFNN, GDFNN and FAOS-PFNN. Simulation results have shown that the proposed algorithm reduces computation time while achieving comparable accuracy.

Surface Finishing and End Milling are among the most sophisticated manufacturing processes. For the industry to improve the quality of its end-line products, it is important for improving the performance of these processes by having a descriptive reference model. Using this reference model, non-intrusive prediction of the resulting surface quality and the tool status can be accurately conducted. Many modeling techniques have been used in literature. Since there are no report of success on a general model that support all the tool specifications, cutting conditions and correlation of the tool-health, cutting signals and resulting surface roughness or tool-wear, the researches based on the several available AI techniques and different sensor signals and there features are on going. This paper investigates the existing correlation between the resulted wavelet coefficients and ball-nose tool-wear using Cascaded Feed Forward Neural Networks (CFFNN). Considering the changes in the shape of the signals during the cutting process and the similarity of the resulting signals to some mother wavelets and the lack of literature on wavelet analysis for ball-nose cutters' signals, this specific analysis is chosen. CFFNN is also selected for its capability to deal with a non-linear process and being comparatively simple. The results are satisfying with the proposed structure. More studies for the optimal structure and features are expected in the future.

This paper proposes a novel illumination normalization approach with low computation complexity for face recognition. In this proposed method, a block-wise Walsh-Hadamard transform (WHT) is employed in the logarithm domain. An appropriate number of low-frequency WHT coefficients are zeroed to compensate for illumination variations. Experiments on different databases demonstrate that the proposed method obtains results comparable to those of conventional Discrete Cosine Transform method but with a higher efficiency. It also achieves better performances for cases with larger illumination variations. Furthermore, both analytical proof and experimental results demonstrate that principal component analysis (PCA) can be directly implemented in the WHT domain.

In this paper, a novel approach towards self-generating fuzzy neural network (SGFNN) is proposed. The proposed approach is simple and effective and is able to generate a fuzzy neural network with high accuracy and compact structure. The structure learning algorithm of the proposed SGFNN combines criteria of rule generation with a pruning technology. The Kalman filter (KF) algorithm is used to adjust the consequent parameters of the SGFNN. The SGFNN is applied for function approximation, nonlinear system identification and time-series prediction problems. Simulation results and comparative studies with other algorithms demonstrate that a more compact architecture with high performance can be obtained by the proposed approach.

In this paper, the viability of using Fuzzy-Rule-Based Regression Modeling (FRM) algorithm for tool performance and degradation detection is investigated. The FRM is developed based on a multi-layered fuzzy-rule-based hybrid system with Multiple Regression Models (MRM) embedded into a fuzzy logic inference engine that employs Self Organizing Maps (SOM) for clustering. The FRM converts a complex nonlinear problem to a simplified linear format in order to further increase the accuracy in prediction and rate of convergence. The efficacy of the proposed FRM is tested through a case study - namely to predict the remaining useful life of a ball nose milling cutter during a dry machining process of hardened tool steel with a hardness of 52-54 HRc. A comparative study is further made between four predictive models using the same set of experimental data. It is shown that the FRM is superior as compared with conventional MRM, Back Propagation Neural Networks (BPNN) and Radial Basis Function Networks (RBFN) in terms of prediction accuracy and learning speed.

In this paper, an online self-organizing scheme for Parsimonious and Accurate Fuzzy Neural Networks (PAFNN), and a novel structure learning algorithm incorporating a pruning strategy into novel growth criteria are presented. The proposed growing procedure without pruning not only simplifies the online learning process but also facilitates the formation of a more parsimonious fuzzy neural network. By virtue of optimal parameter identification, high performance and accuracy can be obtained. The learning phase of the PAFNN involves two stages, namely structure learning and parameter learning. In structure learning, the PAFNN starts with no hidden neurons and parsimoniously generates new hidden units according to the proposed growth criteria as learning proceeds. In parameter learning, parameters in premises and consequents of fuzzy rules, regardless of whether they are newly created or already in existence, are updated by the extended Kalman filter (EKF) method and the linear least squares (LLS) algorithm, respectively. This parameter adjustment paradigm enables optimization of parameters in each learning epoch so that high performance can be achieved. The effectiveness and superiority of the PAFNN paradigm are demonstrated by comparing the proposed method with state-of-the-art methods. Simulation results on various benchmark problems in the areas of function approximation, nonlinear dynamic system identification and chaotic time-series prediction demonstrate that the proposed PAFNN algorithm can achieve more parsimonious network structure, higher approximation accuracy and better generalization simultaneously.

Localization is one of the most important research issues in wireless sensor networks (WSNs). Hop-count based localization has been proposed as a cost-effective alternative to the expensive hardware-based localization schemes. In this paper, we propose a new method to estimate distances between anchors and nodes based on hop-count information. Localization is then achieved using the estimated distances. Simulation results show that the performance of our proposed algorithm is much better than that of the DV-hop algorithm in terms of the node-to- anchor distance estimation and localization accuracy, and the improvement is proportional to the node density.

This paper represents a new approach for automatically generating fuzzy inference system (FIS) using incremental topological preserving map fuzzy Q-learning (ITPM-FQL). The ITPM-FQL can create and tune the fuzzy rules automatically without any priori knowledge. The online self organizing ITPM approach is used to achieve automatic structure identification while the fuzzy Q-learning approach is used for parameter identification to deal with continuous states and actions. Compared with the first authorpsilas previous works in dynamic fuzzy Q-learning (DFQL), this proposed approach is able to achieve fewer numbers of fuzzy rules. Similar to the DFQL, epsiv-completeness criterion is used to generate fuzzy rules but the convergence capability of the ITPM is added to provide flexible fuzzy clustering. Experimental results and comparative studies with conventional fuzzy Q-learning (FQL), continuous-action Q-learning (CAQL), DFQL and its related developments, dynamic self generated fuzzy Q-learning (DSGFQL) and enhanced dynamic self generated fuzzy Q-learning (EDSGFQL), in wall-following task of a mobile robot are presented to demonstrate the superiority of the proposed approach.

In this paper, a self-constructing fuzzy neural networks with extended Kalman filter (SFNNEKF) is proposed. The whole network generalization capability is considered in the hidden neuron growing criterion, which makes the growing process more smoothly. The extended Kalman filter method is used to adjust the free parameters of the fuzzy neural networks. The proposed SFNNEKF learning algorithm is evaluated in channel equalization problems for communication systems. simulation results show that the SFNNEKF equalizer is superior to other equalizers such as recurrent neural network (RNN), minimal resource allocation network (MRAN), the radial basis function neural network (RBFNN) and the growing and pruning RBF (GAP-RBF) network in terms of bit error rate (BER).

IPM(Inverted Pendulum Model) has been widely used for modeling of human motion gaits. There is a common condition in most of these models, the reaction force between the floor and the humanoid must go through the CoG (Center of Gravity) of the a humanoid or human being. However, the recent bio-mechanical studies show that there are angular moments around the CoG of a human being during human motion. In other words, the reaction force does not necessarily pass through the CoG. In this paper, the motion of IPM is analyzed by taking into consideration two kinds of rotational moments, namely around the pivot and around the CoG. The human motion has been decomposed into the sagittal plane and front plane in the double support phase and single support phase. The motions of the IPM in these four different phases are derived by solving four differential equations with boundary conditions. Simulation results show that a stable human gait is synthesized by using our proposed IPM.

Target tracking is one of the very important applications of WSNs (Wireless Sensor Networks). Traditionally, Kalman filter [1] and its derivatives [2, 3] are some of the most popular algorithms in solving the signal tracking problem. In a WSNs tracking application, the target motion/state update dynamics might be linear or nonlinear depending on the specific scenario. The observation model might vary across the sampling interval. This paper compares the effectiveness, limitations and other related implementation issues in applying Kalman filter and extended Kalman filter to tackle target tracking problem in WSNs.

Inherently, the operational processes of a large number of digital communication systems are based on input events. Therefore, discrete event-based system (DEVS) model is an appropriate tool to model such systems. On the other hand, as communication networks become more popular, congestion control becomes more vital. This necessitates the design and development of a suitable model that is able to represent the complete behavior of the network traffic as much as possible. Many researchers have attempted to develop a model first, and then at the next step, develop a good controller for the communication system. In other words, modeling and control are usually treated as two separate issues. However, it seems that treating the modeling and control issues simultaneously may improve the performance and efficiency of communication systems significantly. In other words, the idea is to develop a model which is naturally similar to the system and at the same time is intelligent enough for self-adaptation so that real-time updates of the system's states are possible. General Fuzzy Automata (GFA) is a discrete-event-based formalism and is thus naturally similar to communication systems. Due to the fuzzy nature of GFA, it is possible to incorporate the advantages of fuzzy control. Therefore, in this paper, GFA is proposed as a tool for the modeling and control of communication systems. Simulation studies show that the GFA-based approach is superior to the classical ATM switch in the usage of a system's resources and system performance.

An evolutionary approach toward automatic generation of fuzzy inference systems (FISs), termed evolutionary dynamic self-generated fuzzy inference systems (EDSGFISs), is proposed in this paper. The structure and parameters of an FIS are generated through reinforcement learning, whereas an action set for training the consequents of the FIS is evolved via genetic algorithms (GAs). The proposed EDSGFIS algorithm can automatically create, delete, and adjust fuzzy rules according to the performance of the entire system, as well as evaluation of individual fuzzy rules. Simulation studies on a wall-following task by a mobile robot show that the proposed EDSGFIS approach is superior to other related methods.

Motion planning is an essential task for humanoid robots. However, it is still very challenging to obtain good motion performance in humanoid motion planning, because of its high DOFs (degree of freedoms), variable mechanical structure and nonlinearity. In humanoid motion planning, the motion performance can be given only after one whole cycle motion is completed. This is a demanding condition for motion planning on either real robots or simulation platform. In this paper, a DFNN (dynamic neural fuzzy network) is adopted to model humanoid robots for motion planning. The inputs of DFNN are parameters which determine motion of humanoid robots. The output is evaluation of humanoid motion performance. The DFNN after training can give evaluation of motion performance immediately once the parameters are determined. The DFNN models not only the dynamics of robots, also the motion planning method. Therefore, the DFNN stores two kind of knowledge: the mapping between parameters and humanoid motion, the mapping between humanoid motion and motion performance.

In this paper, a novel approach termed dynamic self-generated fuzzy Q-learning (DSGFQL) for automatically generating fuzzy inference systems (FISs) is presented. The DSGFQL methodology can automatically create, delete and adjust fuzzy rules without any priori knowledge. Compared with conventional fuzzy Q-learning (FQL) approaches which only use reinforcement learning (RL) for the consequents part of an FIS, the most salient feature of the DSGFQL is that it applies RL to generate both preconditioning and consequent parts of the FIS. The preconditioning parts of the FIS are formed by RL approaches as well as the epsiv-completeness criteria. On the other hand, the consequent parts of the FIS are updated by FQL. Compared with our previously proposed generalized dynamic fuzzy neural networks (GDFNN), which is a Supervised Learning (SL) approach, the DSGFQL approach can be applied to situations when the training teacher is not available. Compared with the previously proposed dynamic fuzzy Q-learning (DFQL) and online clustering and Q-value based genetic algorithm learning schemes for fuzzy system design (CQGAF), the DSGFQL approach can delete unsatisfactory and redundant fuzzy rules as well as adjust the membership of fuzzy functions. Simulation studies on a wall-following task by a mobile robot show that the proposed DSGFQL algorithm is superior to DFQL and CQGAF.

System identification is a very important part in control theory for nonlinear analysis and optimization. In the past years, neural identification of dynamic systems gains great interest because of its powerful mapping capability. In this paper, a learning algorithm for the feedforward neural network named extreme learning machine (ELM) is applied for nonlinear system identification problem. The simulation results show that ELM can achieve very satisfying identification performance and fast learning speed

In this paper, a learning control approach is applied to the synchronization of two uncertain chaotic systems which contain nonlinear uncertainties with unknown time delays. This learning approach also deals with unknown time-varying parameters having distinct periods in the master and slave systems. Using the Lyapunov-Krasovskii functional and incorporating periodic parametric learning mechanism, global stability and asymptotic synchronization between the master and the slave systems are obtained. Simulation studies on representative classes of chaotic systems demonstrate the effectiveness of the proposed approach

In this paper, a novel approach termed enhanced dynamic self-generated fuzzy Q-learning (EDSGFQL) for automatically generating a fuzzy inference system (FIS) is presented. In this temporal difference (TD)-based EDSGFQL approach, the structure and preconditioning parts of an FIS are generated by a hybrid reinforcement learning (RL) and unsupervised learning (UL) approach. In the EDSGFQL approach, the preconditioning parts of an FIS are estimated and generated dynamically via evaluations on the TD error and firing strength. An extended self organizing map (SOM) algorithm is used for updating the centers of membership functions (MFs). The consequent parts of the FIS are trained through the FQL. The proposed EDSGFQL methodology can automatically create, delete and adjust fuzzy rules without any priori knowledge, which is superior to many existing methodologies. Simulation studies on a wall-following task by a mobile robot show that the proposed EDSGFQL approach is superior

In this paper, we present a new scheme to design adaptive controller for uncertain nonlinear systems in the presence of input saturation. The control design is achieved by using backstepping technique and neural network. Unlike some existing control schemes for systems with input saturation, the developed controller does not require uncertain parameters within a known compact set. Besides showing stability, transient performance is also established and can be adjusted by tuning certain design parameters

In Turing Machines, the input string can be of size n, where n is any natural number. The memory tape of the Turing Machine may store symbols which in general, may depend on this parameter n. Using hard-limiter neurons, earlier we introduced a framework for simulating a Turing machine [5]. In our earlier schemes [5], for accessing one symbol on memory tape having m symbols, we needed the number of operations that depend (at least logarithmically) on m. In this paper, we introduce a framework for constant time access of any symbol on memory tape.

In this paper, we investigate the iterative learning control of interconnected nonlinear systems with repeatable control tasks. For each subsystem, the main structure of the local learning controller is constructed by a fuzzy learning component and a robust learning component. The learning control algorithm employs fuzzy system and sliding-mode like technique to adaptively compensate for the plant nonlinearities and interconnections. The interaction between each subsystem can be a general type of unknown nonlinear functions. Under a bounding condition on the nonlinear interconnections, we show that all the internal signals are bounded during the learning process and the state tracking errors of each subsystem converge asymptotically along the iteration axis to a tunable residual set.

This paper introduces a novel self-generated gait synthesis approach for dynamic control of Humanoid Robots (HRs). The main idea is to define gait trajectories for hip and ankle so that motions of other joints can be regulated simultaneously. Because stability is one of the most common concerns requested for HRs, a self-learning control strategy of improving dynamic stability based on the zero moment point (ZMP) criterion is developed. As hip motion plays the most important role of dynamic stability, a dynamic fuzzy Q-learning (DFQL) controller is proposed to define the hip motion trajectory. A salient feature of the proposed approach is that the DFQL controller can self generate fuzzy rules without a priori knowledge and it is capable of dealing with dynamic systems. The DFQL controller can automatically generate the structure as well as parameters of the fuzzy system. Simulation results show that the DFQL controller is capable of improving dynamic stability as the actual ZMP trajectory becomes very close to the ideal case.

This paper presents a novel illumination normalization approach for face recognition under varying lighting conditions. In the proposed approach, a discrete cosine transform (DCT) is employed to compensate for illumination variations in the logarithm domain. Since illumination variations mainly lie in the low-frequency band, an appropriate number of DCT coefficients are truncated to minimize variations under different lighting conditions. Experimental results on the Yale B database and CMU PIE database show that the proposed approach improves the performance significantly for the face images with large illumination variations. Moreover, the advantage of our approach is that it does not require any modeling steps and can be easily implemented in a real-time face recognition system.

Memetic algorithms have become to gain increasingly important for solving large scale combinatorial optimization problems. Typically, the extent of the application of local searches in canonical memetic algorithm is based on the principle of "more is better". In the same spirit, the island model parallel memetic algorithm (PMA) is an important extension of the canonical memetic algorithm which applies local searches to every transitional solutions being considered. For PMA which applies complete local search, we termed it as PMA-CLS. In this paper, we consider the island model PMA with selective application of local search (PMA-SLS) and demonstrate its utility in solving complex combinatorial optimization problems, in particular large-scale quadratic assignment problems (QAPs). Based on our empirical results, the PMA-SLS compared to the PMA-CLS, can reduce the computational time spent significantly with little or no lost of solution quality. This we concluded is due mainly to the ability of the PMA-SLS to manage a more desirable diversity profile as the search progresses.

Our goal is to decompose an arbitrary pathwise continuous rectilinear polygon with holes into a fixed number of nonoverlapping pathwise continuous rectilinear open subsets. This goal involves two additional hard constraints – a. The area of each subset must equal an a priori value. b. The union of the closure of the subsets must be equal to the polygon itself. A soft constraint is that any edge that defines the closure of any subset measures ω units or more. We give a decomposition algorithm that satisfies these constraints. Our goal is motivated by a problem involving parallel utilization of multiple unmanned reconnaissance aerial vehicles (URAVs) with heterogeneous endurance characteristics and equipped with similar sensors with heterogeneous capabilities to conduct a photographic scan of regions with urban-like structures. We discuss our motivation in greater details.

An unmanned reconnaissance aerial vehicle mounted sensor of footprint with a small area /spl omega//sub s//sup 2/ is used to cover critical airbase structures for damage assessment. The region of coverage interest is modeled with a closed union of a minimal set of interior disjoint rectangles of width /spl omega//sub i/ = /spl omega//sub s/. We wish to find a minimum-length flight path for complete coverage for the nonholonomic vehicle. We prove that our minimization problem is a traveling salesman problem (TSP) that is symmetric, non-Euclidean and satisfies the triangular inequality. We compare our modeling primitive with two simple primitives commonly used for representing coverage regions. An ant colony optimization heuristic for solving the TSP is presented.

In recent decades, various metaheuristics, such as genetic algorithms (GA), have been proposed to solve the vehicle routing problem (VRP), a well-known class of combinatorial optimization problems. It is generally known that the scheme for genetic representation of the solution albeit the chromosome coding structure, can play a crucial role in GA. Consequently, this may have a profound impact on the algorithm's performance significantly. We propose and study three forms of genetic representations used in a hybrid genetic algorithm for solving the VRP and analyze their influence on the performance of the algorithm. Algorithms with the different solution coding schemes were applied to a test case scenario of supply chain distribution network.

In this paper, a novel adaptive noise cancellation algorithm using enhanced dynamic fuzzy neural networks (EDFNNs) is described. In the proposed algorithm, termed EDFNN learning algorithm, the number of radial basis function (RBF) neurons (fuzzy rules) and input-output space clustering is adaptively determined. Furthermore, the structure of the system and the parameters of the corresponding RBF units are trained online automatically and relatively rapid adaptation is attained. By virtue of the self-organizing mapping (SOM) and the recursive least square error (RLSE) estimator techniques, the proposed algorithm is suitable for real-time applications. Results of simulation studies using different noise sources and noise passage dynamics show that superior performance can be achieved.

In this paper, an efficient method for high-speed face recognition based on the discrete cosine transform (DCT), the Fisher's linear discriminant (FLD) and radial basis function (RBF) neural networks is presented. First, the dimensionality of the original face image is reduced by using the DCT and the large area illumination variations are alleviated by discarding the first few low-frequency DCT coefficients. Next, the truncated DCT coefficient vectors are clustered using the proposed clustering algorithm. This process makes the subsequent FLD more efficient. After implementing the FLD, the most discriminating and invariant facial features are maintained and the training samples are clustered well. As a consequence, further parameter estimation for the RBF neural networks is fulfilled easily which facilitates fast training in the RBF neural networks. Simulation results show that the proposed system achieves excellent performance with high training and recognition speed, high recognition rate as well as very good illumination robustness.

This paper presents an adaptive modeling and control scheme for drug delivery systems based on a generalized fuzzy neural network (G-FNN). The proposed G-FNN is a novel intelligent modeling tool, which can model unknown nonlinearities of complex drug delivery systems and adapt to changes and uncertainties in these systems online. It offers salient features, such as dynamic fuzzy neural topology, fast online learning ability and adaptability. System approximation formulated by the G-FNN is employed in the adaptive controller design for drug infusion in intensive care environment. In particular, this paper investigates automated regulation of mean arterial pressure (MAP) through intravenous infusion of sodium nitroprusside (SNP), which is one attractive application in automation of drug delivery. Simulation studies demonstrate the capability of the proposed approach in estimating the drug's effect and regulating blood pressure at a prescribed level.

This paper is concerned with the design of a sliding mode observer (SMO) for a class of uncertain nonlinear differential-algebraic systems (DAS) described by so-called semi-explicit forms with differential variables being coupled with algebraic variables. In order to estimate the algebraic variables directly, an improved algorithm is developed to reconstruct the algebraic variables that are subject to a singular distribution matrix of algebraic variables, using a series of elementary matrices followed by differentiation. An SMO is then designed based on the reconstructed algebraic variables in order to compensate for the effect of disturbances on estimation error dynamics such that the estimated states, including both the differential and algebraic variables, can follow the actual ones. The stability of the proposed observer is proven and an illustrative example is given to demonstrate the effectiveness of the improved algorithm for constructing the SMO.

This paper presents a novel illumination normalization approach for face recognition under varying lighting conditions. First, we demonstrate that illumination compensation can be efficiently implemented in the logarithm domain. In the proposed approach, discrete cosine transform (DCT) is employed to compensate for illumination variations in the logarithm domain. Since illumination variations mainly lie in the low-frequency band, an appropriate number of DCT coefficients are truncated to reduce the variations under different lighting conditions. The salient feature of our approach is that it does not need any training or modelling step and can be easily implemented with high speed.

In this paper, a dynamic fuzzy Q-learning (DFQL) method navigating a mobile robot efficiently is presented. Self-organizing fuzzy inference is introduced to calculate actions and Q-functions which capable of enabling us to deal with continuous-valued states and actions. Consequently, fuzzy rules can be generated automatically. Fuzzy inference systems provide a natural mean of incorporating the bias components for rapid reinforcement learning. Furthermore, the eligibility trace method is employed in our algorithm, leading to faster learning and alleviating the experimentation-sensitive problem where an arbitrarily bad training policy might result in a non-optimal policy. Experimental results demonstrate that the robot is able to learn the right policy with a few trials.

In this work, a new iterative learning control (ILC) algorithm performing state tracking in the presence of both parametric and non-parametric uncertainties is proposed. To deal with time-varying parametric uncertainties, iterative updating based on the previous iteration's control signal and current iteration's tracking error is employed. For norm-bounded nonparametric uncertainties, iterative updating combined with a robust control scheme is implemented. A kind of energy-function-based approach is utilized for control law design and learning convergence. Rigorous mathematical proof shows that the integration of the two different updating laws and the robust control scheme can guarantee convergence of the proposed iterative learning algorithm.

This paper extends our previous work on the island model parallel hybrid-genetic algorithm (PHGA) for large scale quadratic assignment problems (QAPs). Some issues on the control parameters of the migration process and how they affect the quality of the solutions and the efficiency of algorithm deserve further evaluative study. In this paper, we investigate the effect of migration topology on the performance of the PHGA. Two topologies, one-way ring topology and random topology, are studied and analyzed. The empirical results show that the PHGA with ring topology is better able to achieve an appropriate tradeoff between exploration and exploitation and hence more helpful to improve the performance of PHGA for solving large scale QAPs.

We summarize some of the key research areas in brain sciences and neural networks that have recently been or are being worked on by researchers in Singapore. Researchers in Singapore are developing theory of neural networks, notably improved radial basis function networks, fuzzy neural networks, and fast learning neural networks. Applications of neural networks include bioinformatics, multimedia, data mining, and communications. Researchers are also working with neurophysiologists on functional brain imaging and brain disease analysis.

This paper presents a new approach towards optimal design of a hybrid fuzzy controller for robotics systems. The salient feature of the proposed approach is that it combines the fuzzy gain scheduling method and a fuzzy proportional-integral-derivative (PID) controller to solve the nonlinear control problem. The resultant fuzzy rule base of the proposed controller can be decomposed into two layers. In the upper layer, the gain scheduling method is incorporated with a Takagi-Sugeno (TS) fuzzy logic controller to linearize the robotics system for a given reference trajectory. In the lower layer, a fuzzy PID controller is derived for all the locally linearized systems by replacing the conventional PI controller by a linear fuzzy logic controller, which has different gains for different linearization conditions. Within the guaranteed stability region, the controller gains can be optimally tuned by genetic algorithms. Simulation studies on a pole balancing robot and a multilink robot manipulator demonstrate the effectiveness and robustness of the proposed approach.

This paper presents a dynamic fuzzy Q-learning (DFQL) method that is capable of tuning fuzzy inference systems (FIS) online. A novel online self-organizing learning algorithm is developed so that structure and parameters identification are accomplished automatically and simultaneously based only on Q-learning. Self-organizing fuzzy inference is introduced to calculate actions and Q-functions so as to enable us to deal with continuous-valued states and actions. Fuzzy rules provide a natural mean of incorporating the bias components for rapid reinforcement learning. Experimental results and comparative studies with the fuzzy Q-learning (FQL) and continuous-action Q-learning in the wall-following task of mobile robots demonstrate that the proposed DFQL method is superior.

This paper presents an adaptive modeling and control scheme for blood pressure regulation based on a generalized fuzzy neural network (G-FNN). The proposed G-FNN is a novel intelligent modeling tool, which can model the unknown nonlinearities of complex drug delivery systems and adapt to changes and uncertainties in these systems online. It offers salient features, such as dynamic fuzzy neural topology, fast online learning ability and adaptability, etc. System approximation formulated by the G-FNN is thus employed in the adaptive control of drug infusion for blood pressure regulation. In particular, this paper investigates automated regulation of mean arterial pressure (MAP) through the intravenous infusion of sodium nitroprusside (SNP), which is one of the most attractive applications in automation of drug delivery. Simulation study demonstrates superior performance of the proposed approach for estimating the drug's effect and regulating blood pressure at a prescribed level.

This paper presents a dynamic Q-learning (DFQL) method that is capable of tuning the fuzzy inference systems (FIS) online. On-line self-organizing learning is developed so that structure and parameters identification are accomplished automatically and simultaneously based only on Q-learning. Self-organizing fuzzy inference is introduced to calculate actions and Q-functions so as to enable us to deal with continuous-valued states and actions. Fuzzy rules provide a natural mean to incorporate the bias components for rapid reinforcement learning. Experimental results and comparative studies with the fuzzy Q-learning the wall following task of mobile robots demonstrate the superiority of the proposed DFQL method.

This paper presents an adaptive fuzzy neural control strategy to regulate Mean Arterial Pressure (MAP) through the intravenous infusion of Sodium NitroPrusside (SNP). The proposed indirect adaptive controller involves a feedforward Generalized Fuzzy Neural Network (G-FNN) together with a linear feedback loop. It is capable of achieving real-time fine control under significant uncertainties and without any prior knowledge of the system dynamics. This is achieved through adaptive learning and modeling of the system dynamics and its uncertainties based on the G-FNN. Salient features of the proposed G-FNN include dynamic fuzzy neural structure, fast online learning ability and adaptability, etc. Simulation studies demonstrate the superior performance of the proposed approach for estimating the drug's effect and regulating blood pressure at a prescribed level.

This paper presents a robust adaptive fuzzy neural controller (AFNC) suitable for identification and control of a class of uncertain multiple-input-multiple-output (MIMO) nonlinear systems. The proposed controller has the following salient features: 1) self-organizing fuzzy neural structure, i.e., fuzzy control rules can be generated or deleted automatically; 2) online learning ability of uncertain MIMO nonlinear systems; 3) fast learning speed; 4) fast convergence of tracking errors; 5) adaptive control, where structure and parameters of the AFNC can be self-adaptive in the presence of disturbances to maintain high control performance; 6) robust control, where global stability of the system is established using the Lyapunov approach. Simulation studies on an inverted pendulum and a two-link robot manipulator show that the performance of the proposed controller is superior.

This paper presents a novel approach of controlling a mobile robot using Generalized Dynamic Fuzzy Neural Networks (GDFNN). Using the GDFNN learning algorithm, not only the parameters of the controller can be optimized online, but also the structure of the controller can be self-adaptive. In comparison to the state-of-the-art neuro-fuzzy controller which predefines the rules, the proposed approach is more flexible. Moreover, the learning speed of this approach is very fast and fuzzy rules can be automatically generated online. This is in contrast with the state-of-the-art neuro-fuzzy controller which requires offline learning process. Simulations studies on a Khepera II robot show that the performance of the proposed approach is more superior.

In order to improve the ability of ultrasound (US) in the classification of benign and malignant breast tumors, a survey of shadow effects in B-scan imaging is presented in this paper. In the preliminary work, a set of agar base tissue mimicking phantoms was fabricated with certain regions imitating different levels of backscattering parts. Texture analysis and the proposed feature i.e. brightness of shadow were used to classify phantom composition. Comparison of back-propagation (BP) training errors was made between the classifier with and without consideration of shadow effects. Classification was implemented only on four digitized sonograms due to the lack of real images. Experimental results show that by taking into consideration the shadow effects, the neural networks show a higher rate of convergence and a smaller training error. These findings suggest that shadow effects can be computerized as an additional feature for classification of breast lesions.

Defining a proper fuzzy rule base is the most difficult task in the design of fuzzy controllers. Because there are not much expert knowledge directly available for fuzzy controllers and analytical calculations with nonlinear fuzzy rule bases are very complicated, most fuzzy rule bases defined manually in the literature are linear. In this paper, we propose a recursive fuzzy controller that employs nonlinear fuzzy rules with constraints. The recursive structure and the constraints obtained from analytical calculations guarantee asymptotic stability of the fuzzy controller and the closed-loop system. By using multi-objective evolutionary algorithms for optimal tuning, the proposed controller shows superior performance over a fuzzy controller with linear fuzzy rule base and a conventional state-feedback controller.

The proportional parallel distributed compensation (PPDC) approach is utilized to stabilize time-delay systems modeled by Takagi-Sugeno fuzzy models in this article. Based on the Lyapunov stability analysis, stability conditions concerning asymptotical stability of time-delay systems are established. The main advantage of the PPDC approach over the parallel distributed compensation (PDC) approach is that fewer adjustable parameters are needed to ensure stability. Moreover, the procedure of finding common matrices P and S is simplified and the number of Lyapunov inequalities is reduced significantly. The entire PPDC design procedure employing linear matrix inequalities (LMIs) is presented. A numerical example of stabilizing a continuous stirred tank reactor (CSTR) is given to illustrate salient features of the new approach. © 2002 Wiley Periodicals, Inc.

A general and efficient design approach using a radial basis function (RBF) neural classifier to cope with small training sets of high dimension, which is a problem frequently encountered in face recognition, is presented. In order to avoid overfitting and reduce the computational burden, face features are first extracted by the principal component analysis (PCA) method. Then, the resulting features are further processed by the Fisher's linear discriminant (FLD) technique to acquire lower-dimensional discriminant patterns. A novel paradigm is proposed whereby data information is encapsulated in determining the structure and initial parameters of the RBF neural classifier before learning takes place. A hybrid learning algorithm is used to train the RBF neural networks so that the dimension of the search space is drastically reduced in the gradient paradigm. Simulation results conducted on the ORL database show that the system achieves excellent performance both in terms of error rates of classification and learning efficiency.

This paper presents a new approach to stabilizing a Takagi–Sugeno (TS) fuzzy model. A new controller design called proportional parallel distributed compensation (PPDC) is proposed. Different from other works, PPDC parameters are proportioned, which result in dramatic reduction in control parameters. In our new design, based on the Lyapunov stability analysis, the solution of finding the common positive definite matrix P is simplified and proportional coefficients design can be separated from feedback matrix parameters. Numerical simulation results show that the PPDC's performance is superior to that of PDC. This new approach has promising potential in practical applications of fuzzy systems. © 2001 John Wiley & Sons, Inc.

A fast approach for automatically generating fuzzy rules from sample patterns using generalized dynamic fuzzy neural networks (GD-FNNs) is presented. The GD-FNN is built based on ellipsoidal basis functions and functionally is equivalent to a Takagi-Sugeno-Kang fuzzy system. The salient characteristics of the GD-FNN are: (1) structure identification and parameters estimation are performed automatically and simultaneously without partitioning input space and selecting initial parameters a priori; (2) fuzzy rules can be recruited or deleted dynamically; (3) fuzzy rules can be generated quickly without resorting to the backpropagation (BP) iteration learning, a common approach adopted by many existing methods. The GD-FNN is employed in a wide range of applications ranging from static function approximation and nonlinear system identification to time-varying drug delivery system and multilink robot control. Simulation results demonstrate that a compact and high-performance fuzzy rule-base can be constructed. Comprehensive comparisons with other latest approaches show that the proposed approach is superior in terms of learning efficiency and performance.

A hybrid fuzzy neural networks and genetic algorithm (GA) system is proposed to solve the difficult and challenging problem of constructing a system model from the given input and output data to predict the quality of chemical components of the finished sintering mineral. A bidirectional fuzzy neural network (BFNN) is proposed to represent the fuzzy model and realize the fuzzy inference. The learning process of BFNN is divided into off-line and online learning. In off-line learning, the GA is used to train the BFNN and construct a system model based on the training data. During online operation, the algorithm inherited from the principle of backpropagation is used to adjust the network parameters and improve the system precision in each sampling period. The process of constructing a system model is introduced in details. The results obtained from the actual prediction demonstrate that the performance and capability of the proposed system are superior.

In this paper, an architecture of dynamic fuzzy neural networks (D-FNN) implementing Takagi-Sugeno-Kang (TSK) fuzzy systems based on extended radial basis function (RBF) neural networks is proposed. A novel learning algorithm based on D-FNN is also presented. The salient characteristics of the algorithm are: 1) hierarchical on-line self-organizing learning is used; 2) neurons can be recruited or deleted dynamically according to their significance to the system's performance; and 3) fast learning speed can be achieved. Simulation studies and comprehensive comparisons with some other learning algorithms demonstrate that a more compact structure with higher performance can be achieved by the proposed approach.

A hybrid adaptive fuzzy controller comprises a weighted combination of a direct and an indirect adaptive fuzzy controllers, with a continuously-switched supervisory controller. The direct and indirect adaptive fuzzy controllers allow fuzzy control rules and fuzzy descriptions to be incorporated respectively, to achieve better adaptation speed. The choice of the weights depends on the relative importance and reliability of the available fuzzy control rules and fuzzy descriptions. Theoretical results and simulation studies on a two-link manipulator show that the proposed hybrid adaptive fuzzy controller is robust and stable. It also outperforms previous direct and indirect adaptive fuzzy controllers in terms of tracking accuracy and magnitude of control torque required.