RESUMO
Underwater target detection techniques have been extensively applied to underwater vehicles for marine surveillance, aquaculture, and rescue applications. However, due to complex underwater environments and insufficient training samples, the existing underwater target recognition algorithm accuracy is still unsatisfactory. A long-term effort is essential to improving underwater target detection accuracy. To achieve this goal, in this work, we propose a modified YOLOv5s network, called YOLOv5s-CA network, by embedding a Coordinate Attention (CA) module and a Squeeze-and-Excitation (SE) module, aiming to concentrate more computing power on the target to improve detection accuracy. Based on the existing YOLOv5s network, the number of bottlenecks in the first C3 module was increased from one to three to improve the performance of shallow feature extraction. The CA module was embedded into the C3 modules to improve the attention power focused on the target. The SE layer was added to the output of the C3 modules to strengthen model attention. Experiments on the data of the 2019 China Underwater Robot Competition were conducted, and the results demonstrate that the mean Average Precision (mAP) of the modified YOLOv5s network was increased by 2.4%.
RESUMO
Underwater marine object detection, as one of the most fundamental techniques in the community of marine science and engineering, has been shown to exhibit tremendous potential for exploring the oceans in recent years. It has been widely applied in practical applications, such as monitoring of underwater ecosystems, exploration of natural resources, management of commercial fisheries, etc. However, due to complexity of the underwater environment, characteristics of marine objects, and limitations imposed by exploration equipment, detection performance in terms of speed, accuracy, and robustness can be dramatically degraded when conventional approaches are used. Deep learning has been found to have significant impact on a variety of applications, including marine engineering. In this context, we offer a review of deep learning-based underwater marine object detection techniques. Underwater object detection can be performed by different sensors, such as acoustic sonar or optical cameras. In this paper, we focus on vision-based object detection due to several significant advantages. To facilitate a thorough understanding of this subject, we organize research challenges of vision-based underwater object detection into four categories: image quality degradation, small object detection, poor generalization, and real-time detection. We review recent advances in underwater marine object detection and highlight advantages and disadvantages of existing solutions for each challenge. In addition, we provide a detailed critical examination of the most extensively used datasets. In addition, we present comparative studies with previous reviews, notably those approaches that leverage artificial intelligence, as well as future trends related to this hot topic.
RESUMO
Alzheimer's Disease (AD) is a neurological brain disorder that causes dementia and neurological dysfunction, affecting memory, behavior, and cognition. Deep Learning (DL), a kind of Artificial Intelligence (AI), has paved the way for new AD detection and automation methods. The DL model's prediction accuracy depends on the dataset's size. The DL models lose their accuracy when the dataset has an imbalanced class problem. This study aims to use the deep Convolutional Neural Network (CNN) to develop a reliable and efficient method for identifying Alzheimer's disease using MRI. In this study, we offer a new CNN architecture for diagnosing Alzheimer's disease with a modest number of parameters, making it perfect for training a smaller dataset. This proposed model correctly separates the early stages of Alzheimer's disease and displays class activation patterns on the brain as a heat map. The proposed Detection of Alzheimer's Disease Network (DAD-Net) is developed from scratch to correctly classify the phases of Alzheimer's disease while reducing parameters and computation costs. The Kaggle MRI image dataset has a severe problem with class imbalance. Therefore, we used a synthetic oversampling technique to distribute the image throughout the classes and avoid the problem. Precision, recall, F1-score, Area Under the Curve (AUC), and loss are all used to compare the proposed DAD-Net against DEMENET and CNN Model. For accuracy, AUC, F1-score, precision, and recall, the DAD-Net achieved the following values for evaluation metrics: 99.22%, 99.91%, 99.19%, 99.30%, and 99.14%, respectively. The presented DAD-Net outperforms other state-of-the-art models in all evaluation metrics, according to the simulation results.
Assuntos
Doença de Alzheimer , Humanos , Doença de Alzheimer/diagnóstico por imagem , Inteligência Artificial , Redes Neurais de Computação , Imageamento por Ressonância Magnética/métodos , Encéfalo/diagnóstico por imagemRESUMO
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.
Assuntos
Exoesqueleto Energizado , Lógica Fuzzy , Extremidade Inferior/fisiologia , Processamento de Sinais Assistido por Computador , Algoritmos , Fenômenos Biomecânicos/fisiologia , Desenho de Equipamento , Marcha/fisiologia , Humanos , Aprendizado de Máquina , Dinâmica não Linear , Caminhada/fisiologiaRESUMO
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.
RESUMO
The hemispherical resonator gyroscope (HRG) is a typical capacitive Coriolis vibratory gyroscope whose performance is inevitably influenced by the uneven electrostatic forces caused by the uneven excitation capacitance gap between the resonator and outer base. First, the mechanism of uneven electrostatic forces due to the significantly uneven capacitance gap in that the non-uniformity of the electrostatic forces can cause irregular deformation of the resonator and further affect the performance and precision of the HRG, was analyzed. According to the analyzed influence mechanism, the dynamic output error model of the HRG was established. In this work, the effect of the first four harmonics of the uneven capacitance gap on the HRG was investigated. It turns out that the zero bias and output error, caused by the first harmonic that dominates mainly the amplitude of the uneven capacitance gap, increase approximately linearly with the increase of the amplitude, and periodically vary with the increase of the phase. The effect of the other three harmonics follows the same law, but their amplitudes are one order of magnitude smaller than that of the first one, thus their effects on the HRG can be neglected. The effect of uneven electrostatic forces caused by the first harmonic on the scale factor is that its nonlinearity increases approximately linearly with the increase of the harmonic amplitude, which was analyzed in depth. Considering comprehensively the zero bias, the modification rate of output error, and scale factor nonlinearity, the tolerance towards the uneven excitation capacitance gap was obtained.
RESUMO
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.
RESUMO
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.
RESUMO
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.
Assuntos
Mineração de Dados/métodos , Aprendizado de Máquina , Processamento de Linguagem Natural , Redes Neurais de Computação , AlgoritmosRESUMO
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.
RESUMO
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.
RESUMO
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.
RESUMO
This paper investigates an adaptive sampling rate control scheme for networked control systems (NCSs) subject to packet disordering. The main objectives of the proposed scheme are (a) to avoid heavy packet disordering existing in communication networks and (b) to stabilize NCSs with packet disordering, transmission delay and packet loss. First, a novel sampling rate control algorithm based on statistical characteristics of disordering entropy is proposed; secondly, an augmented closed-loop NCS that consists of a plant, a sampler and a state-feedback controller is transformed into an uncertain and stochastic system, which facilitates the controller design. Then, a sufficient condition for stochastic stability in terms of Linear Matrix Inequalities (LMIs) is given. Moreover, an adaptive tracking controller is designed such that the sampling period tracks a desired sampling period, which represents a significant contribution. Finally, experimental results are given to illustrate the effectiveness and advantages of the proposed scheme.
RESUMO
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.
Assuntos
Lógica Fuzzy , Modelos Teóricos , Redes Neurais de Computação , Navios/classificação , Algoritmos , Simulação por Computador , Movimento (Física)RESUMO
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.
RESUMO
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.
Assuntos
Redes Neurais de Computação , Dinâmica não Linear , HumanosRESUMO
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.
RESUMO
This paper presents a methodology of asymptotically synchronizing two uncertain generalized Lorenz systems via a single continuous composite adaptive fuzzy controller (AFC). To facilitate controller design, the synchronization problem is transformed into the stabilization problem by feedback linearization. To achieve asymptotic tracking performance, a key property of the optimal fuzzy approximation error is exploited by the Mean Value Theorem. The composite AFC, which utilizes both tracking and modeling error feedbacks, is constructed by introducing a series-parallel identification model into an indirect AFC. It is proved that the closed-loop system achieves asymptotic stability under a sufficient gain condition. Furthermore, the proposed approach cannot only synchronize two different chaotic systems but also significantly reduce computational complexity and implemented cost. Simulation studies further demonstrate the effectiveness of the proposed approach.
RESUMO
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.
Assuntos
Algoritmos , Inteligência Artificial , Lógica Fuzzy , Aprendizagem , Redes Neurais de Computação , Pressão Sanguínea/efeitos dos fármacos , Sistemas de Liberação de MedicamentosRESUMO
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.