Search

The interval type-2 Takagi-Sugeno (T-S) fuzzy dynamic output feedback and H-infinity stability analysis is studied for a class of networked control systems with multiple time-varying additive uncertainties, time-varying signal communication delay and external disturbance . Firstly, the interval type-2T-S fuzzy is employed to denote the system plant. Secondly, the multiple time-varying additive uncertainties are introduced in the controller design and the state variables depending on additive uncertainties. Thirdly, the delay-dependent Lyapunov-Krasovskii functional with double integral terms is designed to derive the less conservative stability conditions in terms of linear matrix inequalities (LMIs). The characteristic of the state variables are reflected effectively by employing the controller with multiple time-varying additive uncertainties. The closed-loop system is asymptotically stable with prescribed H-infinity performance index γ by employing the matrices transformation information. The less conservative stability conditions are derived and extended into the networked control system without additive uncertainties. Finally, simulations are presented to show the effectiveness of the proposed methods.

Mobile vehicles require accurate localization and dense mapping for motion planning. In this paper, we propose a dense map construction algorithm based on a light-and-fast stereo visual-inertial navigation system (VINS). A tightly coupled nonlinear optimization method is used to calculate the position of adjacent keyframes. An optical flow tracking method fused with IMU information and ring matching constraints is used to improve the matching accuracy and speed of the feature points. In addition, we obtain the pose and depth values using the semi-global block matching (SGBM) method, which are used as the initial values of the depth filter to update the depth image and improve the convergence speed. Then, we further use the Truncated Signed Distance Function (TSDF) method to construct the dense map. We compare our algorithm with state-of-the-art algorithms on the EuRoc dataset and then compare the estimated depth image using the proposed algorithm and the point cloud construction with the probabilistic monocular dense reconstruction (REMODE). The experiments show that the proposed algorithm can obtain more accurate localization than VINS and OKVIS, as well as a faster tracking speed, a better depth map, a lower convergence time for the estimated image and a lower number of updated frames than REMODE.

Many multi-modal data suffers from significant weak-pairing characteristics, i.e., there is no sample-to-sample correspondence between modalities, rather classes of samples in one modality correspond to classes of samples in the other modality. This provides great challenges for the cross-modal learning for retrieval. In this work, our focus is learning cross-modal representations with minimal class label supervision and without correspondences between samples. To tackle this challenging problem, we establish a scalable hierarchical learning architecture to deal with the extensive weakly-paired heterogeneous multi-modal data. A shared classifier across different modalities is used to effectively deal with label supervision information, and a multi-modal low-rank model is introduced to encourage the modal-invariant representation. Finally, some cross-modal validations on publicly available datasets are performed to show the advantages of the proposed method.

Gestures recognition provides an intelligent, natural, and convenient way for human–robot interaction (HRI). This paper presents a novel data glove for gestures capturing and recognition based on inertial and magnetic measurement units (IMMUs), which are made up of three-axis gyroscopes, three-axis accelerometers and three-axis magnetometers. The proposed data glove has eighteen low-cost IMMUs, which are compact and small enough to wear. The gestures included the three-dimensional motions of arm, palm and fingers are completely captured by the data glove. Meanwhile, we attempt to use extreme learning machine (ELM) for gesture recognition which has not found yet in the relevant application. The ELM-based recognition methods for both static gestures and dynamic gestures are respectively presented. The experimental results of gestures capturing and recognition verify the effectiveness of the proposed methods.

Learning rich representations efficiently plays an important role in the multi-modal recognition task, which is crucial to achieving high generalization performance. To address this problem, in this paper, we propose an effective Multi-Modal Local Receptive Field Extreme Learning Machine (MM-LRF-ELM) structure, while maintaining ELM’s advantages of training efficiency. In this structure, LRF-ELM is first conducted for feature extraction for each modality separately. And then, the shared layer is developed by combining these features from each modality. Finally, the Extreme Learning Machine (ELM) is used as supervised feature classifier for the final decision. Experimental validation on Washington RGB-D Object Dataset illustrates that the proposed multiple modality fusion method achieves better recognition performance.

During the service of the tritium barrier coating, the cracking behavior and the roughness of substrate are important in the fusion reactor. In this study, the extended finite element method (XFEM) is utilized to investigate the relationship between substrate surface roughness and crack behavior in tritium barrier coating system, which consists of Al2O3 coating and 316 L stainless-steel substrate. The results demonstrate that the roughness of the substrate surface affected the strain energy release rate and the propagation patterns. The driving force of crack exhibits different distribution in the peak and valley regions of the substrate surface. The stress, along the path of the rough interface, can be partially suppressed. Moreover, the multiple cracks significantly reduce the growth of cracks.

Based on the triboelectrification and electrostatic induction, triboelectric nanogenerators (TENGs) have already expanded many applications ranging from ambient mechanical energy harvesting to self-powered sensors. Diverse advanced techniques have been utilized to fabricate various devices to fulfill those applications. To further develop the wide utilization of TENGs, we introduce 4D printing technology to fabricate the transparent self-recovered TENGs, which not only provide excellent self-recoverability of device performance and improve the robustness of device structure, but also open a path to fabricate complicated structure through computer design with no need of any molds. The printed devices have the capabilities of harvesting mechanical energy with maximum output power density of 56 mW/m2 as well as detecting the bending angles of human joints as self-powered sensor. The self-recoverability is originated from shape memory polymer (SMP) under thermal treatment. Therefore, the self-recovered TENGs based on 4D printing technology may offer great potential in energy harvesting and self-powered sensors for human-robot cooperation in sensing and control of robot in need of sophisticated and precise structures.

Visual and tactile sensing are complementary factors in the task of robotic grasping. In this paper, a grasp detection deep network is first proposed to detect the grasp rectangle from the visual image， then a new metric using tactile sensing is designed to assess the stability of the grasp. By means of this scheme, a THU grasp dataset， which includes the visual information, corresponding tactile and grasp configurations， is collected to train the proposed deep network. Experiments results have demonstrated that the proposed grasp detection deep networks outperform other mainstream approaches in a public grasp dataset. Furthermore， the grasp success rate can be improved significantly in real world scenarios. The trained model has also been successfully implemented in a new robotic platform to perform the robotic grasping task in a cluttered scenario.

Recognition of complex dynamic texture is a difficult task and captures the attention of the computer vision community for several decades. Essentially the dynamic texture recognition is a multi-class classification problem that has become a real challenge for computer vision and machine learning techniques. Due to the reason that the dynamic textures lie in non-Euclidean manifold, existing classifier such as extreme learning machine cannot effectively deal with this problem. In this paper, we propose a new approach to tackle the dynamic texture recognition problem. First, we utilize the affinity propagation clustering technology to design a codebook, and then construct a soft coding feature to represent the whole dynamic texture sequence. This new coding strategy preserves spatial and temporal characteristics of dynamic texture. Finally, by evaluating the proposed approach on the DynTex dataset, we show the effectiveness of the proposed strategy.

In this paper, a new framework to effectively classify the time sequence is developed. The whole time sequence is divided into several smaller sub-sequence by means of the sliding time window technique. The sub-sequence is modeled as a linear dynamic model by appropriate dimension reduction and the whole time sequence is represented as a bag-of-systems model. Such a model is very flexible to describe time sequence originated from different sensor source. To construct the bag-of-systems model, we design the codebook by using the K-medoids clustering algorithm and Martin distance between linear dynamic systems. Such a technology avoids the problem that linear dynamic systems lie in non-Euclidean manifold. After obtaining the represented of time sequence, an extreme learning machine is utilized for classification. Finally, the proposed method is verified on some benchmark and shows that it obtains promising results.

The random-hidden-node extreme learning machine (ELM) is a much more generalized cluster of single-hidden-layer feed-forward neural networks (SLFNs) which has three parts: random projection, non-linear transformation, and ridge regression (RR) model. Networks with deep architectures have demonstrated state-of-the-art performance in a variety of settings, especially with computer vision tasks. Deep learning algorithms such as stacked autoencoder (SAE) and deep belief network (DBN) are built on learning several levels of representation of the input. Beyond simply learning features by stacking autoencoders (AE), there is a need for increasing its robustness to noise and reinforcing the sparsity of weights to make it easier to discover interesting and prominent features. The sparse AE and denoising AE was hence developed for this purpose. This paper proposes an approach: SSDAE-RR (stacked sparse denoising autoencoder - ridge regression) that effectively integrates the advantages in SAE, sparse AE, denoising AE, and the RR implementation in ELM algorithm. We conducted experimental study on real-world classification (binary and multiclass) and regression problems with different scales among several relevant approaches: SSDAE-RR, ELM, DBN, neural network (NN), and SAE. The performance analysis shows that the SSDAE-RR tends to achieve a better generalization ability on relatively large datasets (large sample size and high dimension) that were not pre-processed for feature abstraction. For 16 out of 18 tested datasets, the performance of SSDAE-RR is more stable than other tested approaches. We also note that the sparsity regularization and denoising mechanism seem to be mandatory for constructing interpretable feature representations. The fact that a SSDAE-RR approach often has a comparable training time to ELM makes it useful in some real applications.

Within the framework of signed graph and multi-agent systems, this paper investigates the distributed bipartite opinion forming problem over coopetition networks. Several sufficient algebraic and geometric topology conditions that guarantee consensus, regardless of the magnitudes of individual coupling strengths among the agents, have been derived by exploring the interaction direction patterns. All the criteria presented do not require the global knowledge of the coupling weights of the entire network, and thus are easier to check. The effectiveness of the theoretical results are illustrated by numerical examples.

This paper proposes a visual teleoperation with human–robot posture-consistent based on deep neural network. A multi-stage structure of visual teleoperation network, in which the angles of robotic joints are obtained from human, is deduced. Furthermore, a novel human–robot posture-consistent mapping method is developed to generate dataset of the visual teleoperation network by solving constrained nonlinear matrix functions. Based on the designed framework, the data generator and a well trained multi-stage visual teleoperation network are presented. Finally teleoperation experiments are implemented to demonstrate that the proposed method is effectiveness and reliable.

At present, the mainstream visible pedestrian detector is easily affected by the ambient lighting, the complex background, and the pedestrians distance. While the infrared images (IR) can compensate for the defects of visible images because of its insensitivity to illumination conditions. Based on the Deep Convolutional Neural Network (DCNN), we proposed a multispectral pedestrian detector that combines visual-optical (VIS) image and infrared (IR) image. We carefully designed three DCNN fusion architectures to study the better fusion stages of the two-branch DCNN. In addition, we compared the three fusion strategies and found that the sum fusion strategy showed better performance to our multispectral detector. Our multispectral pedestrian detectors are more adaptable to the around-the-clock applications such as autonomous driving and unattended monitoring, by testing on the public multispectral benchmark dataset KAIST, our best fusion architectures achieved a log-average miss rate of 27.60% comparable to the state-of-the-art detector, but with half the runtime.

This paper addresses the stability analysis and adaptive robust finite-time bioinspired neurodynamics control for a class of mobile robot systems with unmeasurable angular and linear velocities, and time-varying bounded disturbance. The error system of the mobile robot is decomposed into two subsystems based on the system model. The state feedback control laws with observers are designed for the two subsystems, and the adaptive robust finite-time bioinspired neurodynamics controller (ARFBNC) is designed based on the state feedback control laws and two subsystems. The stability conditions in the form of linear matrix inequalities (LMIs) are derived by introducing the Lyapunov–Krasovskii functional. The unmeasurable angular and linear velocities, and time-varying bounded disturbance are estimated effectively by employing the state feedback control laws with observers. The smooth bounded outputs are obtained and the sharp jumps of initial values for the state variables are reduced. The closed-loop system is asymptotically stable and the state errors converge to an adjustable bounded region by introducing the Lyapunov–Krasovskii functional. The simulations are performed to show the effectiveness of the proposed methods.

The cluster consensus problem of high-order multi-agent systems is considered in this note with an observer-based control scheme. Sufficient condition for cluster consensus is presented in terms of easily checkable algebraic topology criterion. It is found that intra-cluster balanced topologies with antisymmetric (i.e., bidirectional and opposite signed) inter-cluster links promote cluster consensus, irrelevant of the magnitudes of intra-cluster coupling weights. The effectiveness of the theoretical results is illustrated by a numerical example.

In this paper, we investigate action recognition using an inexpensive RGB-D sensor (Microsoft Kinect). First, a depth spatial-temporal descriptor is developed to extract the interested local regions in depth image. Such descriptors are very robust to the illumination and background clutter. Then the intensity spatial-temporal descriptor and the depth spatial-temporal descriptor are combined and feeded into a linear coding framework to get an effective feature vector, which can be used for action classification. Finally, extensive experiments are conducted on a publicly available RGB-D action recognition dataset and the proposed method shows promising results.

In this paper, the theoretical and practical feasibility of estimating ego-motion in monocular navigation is under investigation. In particular, this paper concerns the observability issues of monocular navigation models. In order to recover the ego-motion of the monocular camera from 2D (two-dimension) image sequence, this paper proposes a novel recursive parametric model. This model differs from other traditional parametric models with the reduced state-dimension and the consistent observability of system state. In addition, the consistent observability of different models are compared by using time-varying observability analysis theory. At the end of this paper, two groups of experiments are implemented to validate the effectiveness of the proposed method.

In this paper, we propose a fuzzy partition based T–S fuzzy systems identification method with block structured sparse representation. Firstly, a novel fuzzy partition method is developed to learn fuzzy rule dictionaries by taking advantage of the geometrical structure of input variables and the functional relationship between input and output variables. Then, we explicitly focus on the block structured information existing in T–S fuzzy models and cast the systems identification problem into an optimization problem with structured sparse representation. In such a way, accurate description of T–S fuzzy model is established with far fewer numbers of fuzzy rules by selecting the important fuzzy rules and eliminating the redundant ones in the process of block structured sparse regression. Several numerical experiments on well-known benchmark data sets are carried out to illustrate the effectiveness of the proposed method.

In this paper, robust stochastic stabilization problems for uncertain Markovian jump linear delta operator systems with actuator saturation are considered. The definition of the domain of attraction for a Markovian jump delta operator system in mean square sense is introduced to analyze the stochastic stability of the closed-loop system. By using a stochastic Lyapunov function which is dependent on the jump mode, design procedures for a mode-dependent state feedback controller are developed based on linear matrix inequality approach. A mode-independent state feedback controller is also designed by using a mode-independent Lyapunov function. Some simulation results are given to illustrate the effectiveness of the developed techniques.

Recognizing traffic signs is a challenging problem; and it has captured the attention of the computer vision community for several decades. Essentially, traffic sign recognition is a multi-class classification problem that has become a real challenge for computer vision and machine learning techniques. Although many machine learning approaches are used for traffic sign recognition, they are primarily used for classification, not feature design. Identifying rich features using modern machine learning methods has recently attracted attention and has achieved success in many benchmarks. However these approaches have not been fully implemented in the traffic sign recognition problem. In this paper, we propose a new approach to tackle the traffic sign recognition problem. First, we introduce a new feature learning approach using group sparse coding. The primary goal is to exploit the intrinsic structure of the pre-learned visual codebook. This new coding strategy preserves locality and encourages similar descriptors to share similar sparse representation patterns. Second, we use a non-uniform quantization approach based on log-polar mapping. Using the log-polar mapping of the traffic sign image, rotated and scaled patterns are converted into shifted patterns in the new space. We extract the local descriptors from these patterns to learn the features. Finally, by evaluating the proposed approach using the German Traffic Sign Recognition Benchmark dataset, we show that the proposed coding strategy outperforms existing coding methods and the obtained results are comparable to the state-of-the-art.

Semantic image segmentation is of fundamental importance in a wide variety of computer vision tasks, such as scene understanding, robot navigation and image retrieval, which aims to simultaneously decompose an image into semantically consistent regions. Most of existing works addressed it as structured prediction problem by combining contextual information with low-level cues based on conditional random fields (CRFs), which are often learned by heuristic search based on maximum likelihood estimation. In this paper, we use maximum margin based structural support vector machine (S-SVM) model to combine multiple levels of cues to attenuate the ambiguity of appearance similarity and propose a novel multi-class ranking based global constraint to confine the object classes to be considered when labeling regions within an image. Compared with existing global cues, our method is more balanced between expressive power for heterogeneous regions and the efficiency of searching exponential space of possible label combinations. We then introduce inter-class co-occurrence statistics as pairwise constraints and combine them with the prediction from local and global cues based on S-SVMs framework. This enables the joint inference of labeling within an image for better consistency. We evaluate our algorithm on two challenging datasets which are widely used for semantic segmentation evaluation: MSRC-21 dataset and Stanford Background dataset and experimental results show that we obtain high competitive performance compared with state-of-the-art methods, despite that our model is much simpler and efficient.

This paper describes the neural controller design for the longitudinal dynamics of a generic hypersonic flight vehicle (HFV). The dynamics are transformed into the strict-feedback form. Considering the uncertainty, the neural controller is constructed based on the single-hidden layer feedforward network(SLFN). The hidden node parameters are modified using extreme learning machine (ELM) by assigning random values. Instead of using online sequential learning algorithm (OSLA), the output weight is updated based on the Lyapunov synthesis approach to guarantee the stability of closed-loop system. By estimating the bound of output weight vector, a novel back-stepping design is presented where less online parameters are required to be tuned. The simulation study is presented to show the effectiveness of the proposed control approach.

In this paper, the delay-dependent stability criterion for time-varying delay neural networks in the delta domain is investigated. The unified neural networks, which can be used in both continues-time space and discrete-time space, takes advantage with a high sampling frequency. In the framework of the newly proposed neural networks, the delay-dependent stability criteria is derived in terms of linear matrix inequality by constructing the Lyapunov-Krasovskii function in the delta domain. A numerical simulation is given to show the effectiveness and superiority of the proposed approach.

This paper is concerned with the stabilization problem for a class of discrete-time networked T-S fuzzy systems with bounded time delays and packet losses. By explicitly considering physical properties of networked control systems (NCSs), sufficient conditions for the existence of state feedback fuzzy controller is derived. Then a stabilization approach based on a parallel distributed compensation scheme is developed. The merit of the proposed method lies in its much less conservatism, which is achieved by guaranteeing the deceasement of Lyapunov functional at each control signal updating step. Illustrative examples are provided to show the advantage and effectiveness of the developed results.

This paper investigates the problems of modeling and controller design for complex flexible nonlinear systems (CFNSs) based on a fuzzy singularly perturbed model. A standard discrete-time fuzzy singularly perturbed model (FSPM) is firstly constructed to estimate CFNSs. Based on a matrix spectral norm approach, a static output feedback controller is designed, which guarantees the resulting closed-loop system is asymptotically stable. The gains of controllers are obtained by solving a set of @e-independent linear matrix inequalities (LMIs) so the ill-conditioned problems caused by @e can be easily avoided. Finally, the proposed approach is applied to modeling and attitude control of flexible satellites. In contrast to the existing results, the proposed method has the following advantages: (i) flexible vibrations can be effectively suppressed, and (ii) the control performance is greatly improved, as well as (iii) the control inputs do not need to be limited in controller design. Two practical examples are provided to illustrate the effectiveness of the presented approach. From the simulation results, it can be seen that the proposed approach can also effectively overcome the external disturbances and noises.

In this paper, the direct adaptive neural controller is investigated for the longitudinal dynamics of a generic hypersonic flight vehicle (HFV). The objective of the controller is to make the altitude and velocity to follow a given desired trajectory in the presence of aerodynamic uncertainties. Based on the functional decomposition, the adaptive discrete-time nonlinear controllers are developed using feedback linearization and neural approximation for the two subsystems. Different from the back-stepping design, the altitude subsystem is transformed into the explicit four-step ahead prediction model. With the prediction model, the controller is proposed without virtual controller design. Furthermore, only one direct neural network (NN) is employed for the lumped system uncertainty approximation. The controller is considerably simpler than the ones based on back-stepping scheme and the algorithm needs less NN parameters to be adjusted online. The semiglobal uniform ultimate boundedness (SGUUB) stability is investigated by the discrete-time Lyapunov analysis and the output tracking error is made within a neighborhood of zero. Accordingly, the NN controller is designed for velocity subsystem. The simulation is presented to show the effectiveness of the proposed control approach.

This paper presents a new stabilization method for networked control systems (NCSs). An improved stability criterion dependent on both time delay bound and packet loss bound is derived and the corresponding stabilizing controller design technique is also provided. The merit of the proposed method lies in its much less conservatism, which is achieved by explicitly considering NCSs physical properties and by guaranteeing the deceasement of Lyapunov functional at each control signal updating step rather than at each sampling step, which is largely ignored in the existing literature. Illustrative examples are provided to show the advantage and effectiveness of the developed results.

In this paper, we propose a particle filter that determines the weight of each particle employing the incremental likelihood calculation. Since there is usually a large overlap region between the two particles that are sequentially generated, determining the weight of the particle has only a small time cost. Therefore, the real-time performance of the proposed tracker can be dramatically improved. Extensive experimental results for single-object and multiple-object tracking scenarios are presented to demonstrate the efficiency of the proposed approach. Finally, an interesting color-based active vision system is developed for a free-floating space robot testbed to facilitate teleoperation.

Failure diagnosis and detection of fuzzy discrete event systems play a significant role in the study of complex systems. In this paper, we investigate the diagnosability of fuzzy discrete event systems by proposing a new algorithm based on the concept of undistinguishable strings. Moreover, a necessary and sufficient condition for fuzzy diagnosability is obtained in terms of certain properties of the diagnoser, which is constructed with respect to the minimal observable event. The computing process to check the diagnosability of fuzzy DESs and some examples serving to illuminate the applications are developed and described.

This paper studies adaptive attitude synchronization of spacecraft formation with possible time delay. By introducing a novel adaptive control architecture, decentralized controllers are developed, which allow for parameter uncertainties and unknown external disturbances. Based upon graph theory, Lyapunov stability theory and time-delay control theory, analytical tools are also provided. A distinctive feature of this work is to address the adaptive attitude synchronization with unknown parameters and coupling time delay in a unified theoretical framework, with general directed information flow. It is shown that arbitrary desired attitude tracking and synchronization with respect to a given reference can be attained. Simulation results are provided to demonstrate the effectiveness of the obtained results.

Various physical parameters, including gas concentrations (O2, CO, CH4, and H2) and temperatures at different air velocities, were determined for full scale wood fires in the Chongqing Coal Research Institute fire test tunnel. Both experimental measurements and numerical simulations are discussed. The numerical analysis was performed with the computational fluid dynamics software package “FLUENT”. The results show that the experimental data agree with the simulation results. The results verify that Roberts’ theory of burning is correct. They also prove that the air velocity is the key factor that determines the type of combustion. Also, it is shown that secondary disasters are unlikely for oxygen rich combustion with a limited fire load.

In this paper, a new operator is proposed to optimize the traditional Hopfield neural network (HNN). The key idea is to incorporate the global search capability of the Estimation of Distribution Algorithms (EDAs) into the HNN, which typically has a powerful local search capability and fast operation. On account of this property of the EDA, our proposed algorithm also exhibits a powerful global search capability. In addition, the possible infeasible solutions generated during the re-sampling period of the EDA are eliminated by the HNN. Therefore, the merits of both these methods are combined in a unified framework. The proposed model is tested on a numerical example, the max-cut problem. The new and optimized model yielded a better performance than certain traditional intelligent optimization methods, such as HNN, genetic algorithm (GA). The proposed mutation Hopfield neural network (MHNN) is also used to solve a practical problem, aircraft landing scheduling (ALS). Compared with first-come-first-served sequence, MHNN sequence reduces both total landing time and total delay.

This paper focuses on tracking control problem of flexible-link manipulators. In order to alleviate the effects of nonlinearities and uncertainties, a combined control strategy based on neural network (NN) and the concept of sliding mode control (SMC) is proposed systematically. The chattering phenomenon in conventional SMC is eliminated by incorporated a saturation function in the proposed controller, and the computation burden caused by model dynamics is reduced by applying a two-layer NN with an analytical approximated upper bound, which is used to implement a certain functional estimate. In addition, the Lyapunov analysis can guarantee the signals of closed-loop system bounded and the online NN adaptive laws can make the system states converge to the sliding surface. Furthermore, the boundary layer thickness as well as the gain of corrective control term is also discussed in detail. At last, the theoretic results are validated on the flexible-link manipulator experimental system in Tsinghua University.

In this paper, the stochastic stabilization problem for a class of Markov jumping linear systems (MJLS) subject to actuator saturation is considered. The concept of domain of attraction in mean square sense is used to analyze the closed-loop stability. When the jumping mode is available, a mode-dependent state feedback controller is developed. Otherwise, we give a less conservative approach to design the mode-independent state feedback controller. Both design procedures can be converted into a set of linear matrix inequalities (LMIs). Finally, a numerical example is provided to show the effectiveness of the techniques.

In this paper, both the state feedback and static output feedback H^~ controllers for fuzzy singularly perturbed systems are investigated. We give sufficient conditions for the existence of H^~ controller such that the closed-loop fuzzy control system is globally stable and achieves a prescribed level of disturbance attenuation for sufficiently small perturbation parameters. The linear matrix inequality (LMI) approach is proposed to obtain the state-feedback gains, and a homotopy-based iterative LMI algorithm is developed to get the static output feedback gains. Simulation results for two examples are included to demonstrate the effectiveness of the proposed approaches.

This paper addresses the reduced-order H∞ filtering problem for continuous-time Makovian jump linear systems, where the jump parameters are modelled by a discrete-time Markov process. Sufficient conditions for the existence of the reduced-order H∞ filter are proposed in terms of linear matrix inequalities (LMIs) and a coupling non-convex matrix rank constraint. In particular, the sufficient conditions for the existence of the zero-order H∞ filter can be expressed in terms of a set of strict LMIs. The explicit parameterization of the desired filter is also given. Finally, a numerical example is given to illustrate the proposed approach.

In the above paper, Chang et al. (Fuzzy Sets and Systems 133 (2003) 37) proposed an interesting approach to design the fuzzy static output feedback controller with a common observability Gramian. However, since it involves a nonlinear output function, the results obtained are not correct for the general case. This note intends to circumvent these problems by modifying the approach proposed in Chang et al. (Fuzzy Sets and Systems 133 (2003) 37).

This paper presents a stable neuro-fuzzy (NF) adaptive control approach for the trajectory tracking of the robotic manipulator with poorly known dynamics. Firstly, the fuzzy dynamic model of the manipulator is established using the Takagi-Sugeno (T-S) fuzzy framework with both structure and parameters identified through input/output data from the robot control process. Secondly, based on the derived fuzzy dynamics of the robotic manipulator, the dynamic NF adaptive controller is developed to improve the system performance by adaptively modifying the fuzzy model parameters on-line. The dynamic NF system aims to approximate the whole robot dynamics rather than its nonlinear components as is done by static neural networks. The dynamic inversion introduced for the controller design is constructed by the dynamic NF system and will help the NF controller design because it does not require the assumption that the robot states should be within a compact set. It is generally known that the compact set cannot be specified before the control loop is closed. Thirdly, the system stability and the convergence of tracking errors are guaranteed by Lyapunov stability theory, and the learning algorithm for the dynamic NF system is obtained thereby. Finally, simulation studies are carried out to show the viability and effectiveness of the proposed control approach.

A stable discrete time adaptive control approach using dynamic neural networks (DNNs) is developed in this paper for the trajectory tracking of a robotic manipulator with unknown nonlinear dynamics. By using dynamic inversion constructed by a DNN, the assumption under which the system state should be on a compact set can be removed. This assumption is usually required in neuro-adaptive control. The NN-based variable structure control is designed to guarantee the stability and improve the dynamic performance of the closed-loop system. The proposed control scheme ensures the global stability and desired tracking as well.

A neuro-fuzzy (NF)-based adaptive controller is presented in this paper for the trajectory tracking control of a flexible-link space manipulator. Based on the singular perturbation method and two time-scale decompositions, the dynamic model of a flexible space robot is reduced into a slow subsystem of an equivalent rigid-link arm and a fast subsystem of flexible mode. A dynamic NF adaptive controller based on dynamic inversion is presented for the tracking control of the equivalent rigid arm, while a fuzzy proportional-derivative (PD) controller is used to stabilize the elastic dynamics. Finally, experiment studies are carried out to show the viability and effectiveness of the proposed control approach.

A neural network (NN)-based adaptive control law is proposed for the tracking control of an n-link robot manipulator with unknown dynamic nonlinearities. Basis function-like networks are employed to approximate the plant nonlinearities, and the bound on the NN reconstruction error is assumed to be unknown. The proposed NN-based adaptive control approach integrates the NN approach and an adaptive discrete variable structure control with a simple estimation mechanism for the upper bound on the NN reconstruction errors, and additional control input as a function of the estimate. Lyapunov stability theory is used to prove the uniform ultimate boundedness of the tracking error.