Search

In this paper, a Lyapunov-based fuzzy logic control (FLC) system is developed for controlling of complicated as well as distinct nonlinear systems. According to Lyapunov Stability Theory, a candidate Lyapunov function with simple quadratic form is designed. Via constructing the fuzzy IF–THEN rules through referencing the statuses of errors states and error derivatives in each sub-space of the Lyapunov function derivatives, the proposed structure of FLC system is able to appropriately as well as flexibly adjust the control forces with minimum magnitude compensating the nonlinear systems in real-time. In addition, the designed FLC system can be applied to different kinds of control systems without further information and operation experience. Two nonlinear systems with distinct structures, classical Lorentz system and Mathieu-van der Pol system, are illustrated for simulation examples. In comparison of the previous research work, the simulation results reveal the effectiveness, flexibility and convenient-design of the proposed method.

The state of the art of mobile robot path planning is composed with global planner and local planner. For example, the global planner majorly establishes a suitable path from a knowing map based on the shortest path. And the local planner which computes the velocity command that includes obstacle clearance and progress toward the goal. However, due to the 2D laser ranger on a mobile service robot just detects the part of the meal table or misses it, this will cause robot often hits the meal table. In this work we apply adaptive control method to feed the goals for Dynamic Window Approach’s (DWA) computing. The experimental result shows that DWA will follow the pre-defined path closely and smoothly.

Most SSVEP-based stimuli BCIs are pre-defined using the white blocks. This kind of scenario lead less flexibility in the real life. To represent the flickers with the location, types and configurations of the objects in real world, this paper proposes an SSVEP-based BCI using real-time camera view with object recognition algorithm to provide intuitive BCI for users. A deep learning-based object recognition algorithm is used to calculate the location of the objects on the online camera view from a depth camera. After the bounding box of the objects is estimated, the location of the SSVEP flickers are designed to overlap on the object locations. An overlapping FFT and SVM is used to recognize the EEG signals into corresponding classes. In experimental results, the classification rate for camera view scenario is more than 94.1%. The results show that proposed SSVEP stimuli design is available to create an intuitive and reliable human machine interaction. The proposed results can be used for the users who have motor disabilities to further used to interact with assistive devices, such as: robotic arm and wheelchairs.

This paper proposes a real-time coplanar Non-Uniform Rational B-Spline (NURBS) curve fitting in a CAD/CAM system to improve fitting quality and reduce calculation effort. Curve segmentation is applied to filter the line segments and reshape the different NURBS curve segments. Both position errors and the first derivative are considered to avoid over fitting of the NURBS curve. After the fitted NURBS curve is obtained, the corresponding interpolation for each interpolated time is further processed to obtain the motion command for each axis. The real-time NURBS curve fitting process consists of (1) discrete points import, (2) linear segment filter, (3) curve reshaping, (4) curve fitting considering both the position errors and the first derivative. Simulation results show the proposed approach effectively improves the fitting quality and calculation efficiency.

This paper proposed a novel intuitive teaching technologies by reconstructing the recorded motion information during human guided robotic arm. A learning algorithm is proposed in this paper to recognize the motion primitives according to therbligs definition. The hybrid sensing interface is used to record and modified the positional trajectory, force/torque, and gripper information. Furthermore, an extended Kalman filter is used to pre-process the data and estimate the velocity and acceleration profile as motion features. The motion features, output data via the hybrid sensing interface, is finally used to recognize the target therblig by proposed cascade support vector machine. The experimental results show that the proposed method can recognize the motion features into therbligs correctly and efficiently. The recognition results can be further used to reconstruct an assembly operation.

In this paper, a set of novel smart timely-monitoring fault-diagnosis method is designed, which utilize the characteristics and mapping of the chaos system, provides a high efficiency and accurate identification strategy for the normal and abnormal states of rolling-element bearing. In the past, numerous techniques and approaches were first applied to the data recorded in the running object, and then the features were manually extracted and selected in raw data, and in order to obtain the features accurately, which had to rely on the domain-knowledge and experience of each line, while statistical methods were also necessary for data analysis. In this paper, a smart machine with feature-production is designed, which requires to be mathematically chaotic mapped, in order to reveal the key-features for experimental data. In addition, In addition, by applying the appropriate feature productions and extractions methods built in this paper, such as: Euclidean Feature Value (EFV), K-means Feature Value (KFV), etc., can be timely-monitored of normal and abnormal range feature-values (FV), and applied to the rolling-element bearing system for smart fault diagnosis. The experimental results show that the proposed method has a practical level of performance, and can achieve a high accurate classification rate between normal and abnormal states.

Image matching is a practical tool in the computer vision domain. 2D and 3D image matching can be based on (2D) features or on view and point (3D) which can be cloud based. Intensive practical research has been done on both 2D and 3D image matching and the methods are generally designed to meet the requirements of industrial alignment applications and include such aspects as full rotation, scaling and the handling of multiple objects. The aim of this study is the presentation of a review of 2D and 3D image matching and the provision of a comprehensive reference source for those involved in image matching.

Display Omitted We propose an accelerating CPU based correlation-based image alignment.The image pyramid search scheme is combined with the parallel computation.Sub-pixel accuracy is used to attain the more accurate image alignment.It can align the accurate pose of the template image within the inspected image. This paper proposes an accelerating correlation-based image alignment using CPUs for time-critical applications in automatic optical inspection (AOI). In order to improve computation efficiency, the image pyramid search scheme is combined with the parallel computation. The image pyramid search scheme is employed first to quickly find certain objects in both monochrome and color images with rotation, translation and scaling. Sub-pixel accuracy is then used to attain the more accurate results at the sub-pixel level. In our experimental results, rotation accuracy is smaller than 0.218?, and the speed is increased between 277 and 20,841 times. According to translation, rotation and scaling tests, the errors of rotation, translation and scaling are 0.2?, 2.07pixel and 0.55%, respectively. These results show that the proposed method is suitable for dealing with correlation-based image alignment for time-critical applications in automatic optical inspection.

Recently, the need of pedestrian detection at night has gained more and more interest. However, the performance of traditional nighttime pedestrian detection systems remains poor because region of interest (ROI) generation and feature extraction are designed separately. Thus, this paper presents novel thermal imagery algorithms to enhance the performance of nighttime pedestrian detection systems. The proposed thermal image pedestrian detection system involves novel outline features, developed from the ROI generation method of pedestrians that are different from traditional features. A three-layer back-propagation feed-forward neural network is used as the classifier. Two databases, the OTCBVS database and our own are used to evaluate the performance of the proposed thermal image pedestrian detection algorithm. Experimental results show that the proposed outline features are effective, and the detection performance of a traditional pedestrian detection system at night is improved. A pioneering thermal imagery algorithm for pedestrian detection system is presented.The proposed pioneering outline features are developed from the ROI generation.Experimental results show that the presented features are significantly effective.

This paper proposes an optimal method for NCC-based template matching on modern CPUs for time-critical applications. In order to achieve the superior computation efficiency, the brand-and-bound (BB) scheme and the streaming SIMD extensions 2 (SSE2) instructions are employed to quickly find out the target object with rotation, translation and scaling in monochrome or color image. And we show how to reject unpromising image location very quickly using BB scheme in search process. Furthermore, an efficient implementation for similarity coefficient calculation is also pointed out by using the integration SSE2 instructions. Finally, the results show that the proposed method is very powerful when dealing with the NCC-based template matching in monochrome and color images.

To effectively improve treatment precision in conformal radiotherapy, ultrasound sensing is employed to detect tumors in real time. In order to implement this new configuration, the physician operating the ultrasound probe must not stay in the radiotherapy room. This motivates the development of a physician-commanded robotic manipulation system for target localization in conformal radiotherapy. The proposed robotic control system is in eye-to-hand configuration with binocular vision and force sensing. The ultrasound probe mounted on the end-effector of the manipulator is controlled to follow the pose commanded by the physician. Based on the measurements from a six-axis force sensor, the probe can be controlled further to maintain the desired contact force against the possibly-deforming treatment surface. The proposed hybrid force and vision-based control approach has been validated by experimenting with a six-degrees-of-freedom industrial manipulator to allow clear ultrasound images.

This paper represents a template matching using statistical model and parametric template for multi-template. This algorithm consists of two phases: training and matching phases. In the training phase, the statistical model created by principal component analysis method (PCA) can be used to synthesize multi-template. The advantage of PCA is to reduce the variances of multi-template. In the matching phase, the normalized cross correlation (NCC) is employed to find the candidates in inspection images. The relationship between image block and multi-template is built to use parametric template method. Results show that the proposed method is more efficient than the conventional template matching and parametric template. Furthermore, the proposed method is more robust than conventional template method.

Image registration is wildly used in the biomedical image, but there are too many textures and noises in the biomedical image to get a precise image registration. In order to get the excellent registration performance, it needs more complex image processing, and it will spend expensive computation cost. For the real time issue, this paper proposes edge gradient direction image registration applied to Computer Tomography(CT) image and Ultrasonography (US) image based on the cooperation of Graphic Processor Unit (GPU) and Central Processor Unit (CPU). GPU can significantly reduce the computation time. First, the CT image slice is extracted from the CT volume by the region growing and the interpolation algorithm. Secondly, the image pre-processing is employed to reduce the image noises and enhance the image features. There are two kinds of the image pre-processing algorithms invoked in this paper: 1) median filtering and 2) anisotropic diffusion. Last but not least, the image edge gradient information is obtained by Canny operator, and the similarity measurement based on gradient direction is employed to evaluate the similarity between the CT and the US images. The experimental results show that the proposed architecture can distinctively improve the efficiency and are more suitably applied to the real world.

This paper presents a corner-based image alignment algorithm based on the procedures of corner-based template matching and geometric parameter estimation. This algorithm consists of two stages: 1) training phase, and 2) matching phase. In the training phase, a corner detection algorithm is used to extract the corners. These corners are then used to build the pyramid images. In the matching phase, the corners are obtained using the same corner detection algorithm. The similarity measure is then determined by the differences of gradient vector between the corners obtained in the template image and the inspection image, respectively. A parabolic function is further applied to evaluate the geometric relationship between the template and the inspection images. Results show that the corner-based template matching outperforms the original edge-based template matching in efficiency, and both of them are robust against non-liner light changes. The accuracy and precision of the corner-based image alignment are competitive to that of edge-based image alignment under the same environment. In practice, the proposed algorithm demonstrates its precision, efficiency and robustness in image alignment for real world applications.

The laser manufacture has been widely used in micro-machining as a burgeoning technology. Galvanometric scanning systems is an important component in laser machines, but the galvanometric scanning system is usually combines the image field distortion that is from the mechanism and optical devices. The paper has presented a novel correction algorithm based on ride regression method for compensates the image field distortion. And the Euclidean distance error and the maximum error are 0.0422 and 0.09, respectively. This novel correction technique can effectively increase the accuracy of laser spot position and can further improve the performance of the galvanometric scanning system.

This paper focuses on the guaranteed cost stability analysis of fuzzy-model-based (FMB) control systems. Representing the nonlinear plant using a Takagi–Sugeno (T–S) fuzzy model, a fuzzy controller is employed to close the feedback loop. A weighted linear quadratic cost function is considered as the cost index to measure the performance of the closed-loop fuzzy system in terms of the system states, system outputs, and control signals. The stability of the FMB control system is investigated by the Lyapunov stability theory subject to the minimization of cost index for performance realization. A membership-function-dependent approach using the piecewise-linear membership functions is employed to include the information of membership functions into the stability analysis. Membership-function-dependent stability conditions in terms of linear matrix inequalities are obtained to determine the system stability and feedback gains with the consideration of the system performance measured by the cost function. A simulation example is provided to illustrate the effectiveness and merits of the proposed approach.

A three-degree-of-freedom (3-DOF) dynamic model-based intelligent nonsingular terminal sliding mode control (INTSMC) system is proposed in this study for the precision contours tracking of a gantry position stage. A Lagrangian equation-based 3-DOF dynamic model for the gantry position stage is derived first. Then, to minimize the synchronous error and tracking error in the precision contours tracking, the 3-DOF dynamic model-based INTSMC system is proposed. In this approach, a nonsingular terminal sliding mode control is designed for the gantry position stage to achieve finite time tracking control. Moreover, to increase the robustness and to improve the control performance, an interval type-2 recurrent fuzzy neural network, and asymmetric membership function, which combines the advantages of interval type-2 fuzzy logic system, recurrent neural network, and asymmetric membership function, is developed as an estimator to approximate a lumped uncertainty. Finally, some experimental results of the gantry position stage for optical inspection application are obtained to show the validity of the proposed control approach.

This paper proposes a robust fault diagnosis system of rotating machine adapting machine learning technology. The kernel of this diagnosis system includes a set of individual neural networks based on structured genetic algorithm (sGAINNs). First, the frequency characteristics from differential signals, including fast Fourier transform (FFT) and full spectrum, are used to feed into the sGAINNs corresponding to assigned faults to emphasize the phenomenon of each fault. Especially, the structured genetic algorithm is applied to get the optimal parameters of the above sGAINNs. In the final step of proposed diagnosis system, the evaluated indexes from sGAINNs are synthesized by a reasoning engine to identify the faults in the rotor system. Finally, six common faults of rotor system, unbalance, bow, misalignment, rub, whirl, and whip, are generated from a rotor kit, produced by Bently Nevada Corporation, to verify the performance of this diagnosis system. The advantage of this diagnosis system is that the optimal sGAINNs parameters can be automatically obtained, the local optimal solutions can be reduced and the diagnosis accuracy can be improved.

This paper presents a Fourier descriptor based image alignment algorithm (FDBIA) for applications of automatic optical inspection (AOI) performed in real-time environment. It deliberates component detection and contour tracing algorithms and uses the magnitude and phase information of Fourier descriptors to establish correspondences between the target objects detected in the reference and the inspected images, so the parameters for aligning the two images can be estimated accordingly. To enhance the computational efficiency, the proposed component detection and contour tracing algorithms use the run length encoding (RLE) and Blobs tables to represent the pixel information in the regions of interest. The Fourier descriptors derived from the component boundaries are used to match the target objects. Finally, the transformation parameters for aligning the inspected image with the reference image are estimated based on a novel phase-shifted technique. Experimental results show that the proposed FDBIA algorithm sustains similar accuracy as achieved by the commercial software Easyfind against various rotation and translation conditions. Also, the computational time consumed by the FDBIA algorithm is significantly shorter than that by Easyfind.

This paper proposes a real-time coplanar Non-Uniform Rational B-Spline (NURBS) curves fitting and interpolation for 6-DOF robot arm to improve the fitting and interpolation quality, and reduce the calculation effort. The process consists of (1) reading discrete points, (2) discrete points extraction, (3) curve fitting, (4) interpolation, and (5) inverse kinematics transformation. According to the simulation and experimental results, the proposed approach certainly raises the fitting and interpolation quality, and the efficiency of calculation.

The existing trademark image retrieval (TIR) approaches mostly use complex image features, the integration of multi features, a tree structure, etc. to enable highly accurate retrieval. However, there is the heavy computational burden for complex image features and maximum similarity subtree isomorphism (MSSI) measurement. This paper aims to provide an efficient solution for TIR in real-time applications, especially in measuring the similarity between multi-object trademark images. In particular, we propose a novel algorithm for tree similarity measurement based on the fuzzy inference system (FIS) to improve retrieval efficiency. Furthermore, the integration of global and local geometric descriptors is used to enable accurate retrieval. The global descriptor is computed by employing the Hu moments, while the local descriptors are generated by using a tree structure based on the five geometric features: convexity, eccentricity, compactness, circle variance, and elliptic variance. During the retrieval process, the similarity coefficient between the query and the database image is obtained from the similarity of the global and local descriptors. The proposed technique is evaluated using 1800 trademark images, including 12 different classes and 416 trademark images. Additionally, the three common indices, the precision/recall rate, the Bull’s eye score, and the average normalized modified retrieval rank (ANMRR) are used as the performance indices. The experimental results show that the proposed technique is superior to the other two competitive approaches. It shows 19.43% and 26.78% precision/recall improvement, 19.56% and 30.58% improvement in the average Bull’s eye score, and 0.167 and 0.236 improvement in the ANMRR score, respectively, for the 416 query images. It can be concluded from the experimental analysis that the proposed technique not only provides reliable retrieval results but also improves the retrieval efficiency by 151 times in the retrieval process.

In this article, a neural network corrector is proposed to correct the image shift, yielding the degradation of three-dimensional image reconstruction, for each slice captured by cone-beam computed tomography simulator. There are 3 degrees of freedom in tube module of simulator; the central point of tube module should be aligned with the central point of detector module to guarantee the accurate image projection. However, the mechanism manufacturing and assembling tolerance will let the above aim cannot be met. Here, a standard kit is made to measure the image shift by 1° step from −10° to 10°. The measure data will be the input training data of proposed neural network corrector, and the corrected translation position will be the output of neural network corrector. The Levenberg–Marquardt learning algorithm adjusts the connected weights and biases of the neural network using a supervised gradient descent method, such that the defined error function can be minimized. To avoid the problem of overfitting and improve the generalized ability of the neural network, Bayesian regularization is added to the Levenberg–Marquardt learning algorithm. After the training of neural network corrector, the different target position commands are fed into the neural network corrector. Then, the corrected data from neural network corrector are fed to be the new position command to verify the image correction performance. Moreover, a phantom kit is made to check the corrected performance of the neural network corrector. Finally, the experimental results verify that the image shift can be reduced by the neural network corrector.

Wen-Chung Chang,1,* Chin-Sheng Chen,2,* Hung-Chi Tai,3 Chia-Yuan Liu,4,5 Yu-Jen Chen3 1Department of Electrical Engineering, National Taipei University of Technology, Taipei, Taiwan; 2Graduate Institute of Automation Technology, National Taipei University of Technology, Taipei, Taiwan; 3Department of Radiation Oncology, Mackay Memorial Hospital, Taipei, Taiwan; 4Department of Internal Medicine, Mackay Memorial Hospital, Taipei, Taiwan; 5Department of Medicine, Mackay Medical College, New Taipei City, Taiwan *These authors contributed equally to this work Abstract: The current practice of radiotherapy examines target coverage solely from digitally reconstructed beam's eye view (BEV) in a way that is indirectly accessible and that is not in real time. We aimed to visualize treatment targets in real time from each BEV. The image data of phantom or patients from ultrasound (US) and computed tomography (CT) scans were captured to perform image registration. We integrated US, CT, US/CT image registration, robotic manipulation of US, a radiation treatment planning system, and a linear accelerator to constitute an innovative target visualization system. The performance of this algorithm segmented the target organ in CT images, transformed and reconstructed US images to match each orientation, and generated image registration in real time mode with acceptable accuracy. This image transformation allowed physicians to visualize the CT image-reconstructed target via a US probe outside the BEV that was non-coplanar to the beam's plane. It allowed the physicians to remotely control the US probe that was equipped on a robotic arm to dynamically trace and real time monitor the coverage of the target within the BEV during a simulated beam-on situation. This target visualization system may provide a direct remotely accessible and real time way to visualize, verify, and ensure tumor targeting during radiotherapy. Keywords: ultrasound, computerized tomography, robotic arm, real time target visualization

This paper presents a hierarchical classification of Optical Character Recognition (OCR) using global and local features with modified k-means and neural network. A hierarchical classification strategy, which contains training and recognition phases, was designed to precisely recognize the characters. In the training phase, the global features of sample characters and modified k-means clustering are employed to quickly categorize the characters into several clusters. Then each cluster and its corresponding local features of sample characters are fed into the Back-Propagation Neural Network (BPN) to learn the optimal weights. In the recognition phase, both global and local features are extracted from the test characters. Then the character is recognized by well-trained clusters’ center and neural network for coarse and fine recognition, respectively. Experimental results of the three different datasets tested showed classification rates of 97.22%, 100% and 90.28%, respectively.

A corner-based image alignment algorithm based on the procedures of corner-based template matching is presented in this study. This algorithm consists of two stages: training and matching. In the matching phase, the corners are obtained using Harris corner detection algorithm. These corners are then used to build the pyramid images. In the matching phase, the corners are obtained using the same corner detection algorithm. The similarity measure is then determined by the differences of gradient vector between the corners obtained in the template image and the inspection image, respectively. Furthermore, it further applied the refined function to evaluate the geometric relationship between the template and the inspection images. Results show that the corner-based template matching outperforms the original edge-based template matching in efficiency, and both of them are robust against lighting changes.

In this paper, a novel clustering method which combines the advantages of the density-based algorithm for discovering clusters in large spatial databases with noise (DBSCAN) and K-means is proposed. The proposed method can classify the pathological cell and the normal cell to two cluster memberships and the disturbances can also be eliminated from the image. In addition, by morphological image processing and Sobel edge algorithm, the pathological cell image can be segmented accurately from the image with pathological cell and normal cell. Finally, some images are illustrated to demonstrate the proposed method is superior to K-means and DBSCAN.

This paper proposes a novel image alignment algorithm based on rotation-discriminating ring-shifted projection for automatic optical inspection. This new algorithm not only identifies the location of the template image within an inspection image but also provides precise rotation information during the template-matching process by using a novel rotation estimation scheme, the so-called ring-shifted technique. We use a two stage framework with an image pyramid searching technique for realizing the proposed image alignment algorithm; in the first stage, the similarity based on hybrid projection transformation with the image pyramid searching technique is employed for quick selection and location of the candidates in the inspection image. In the second stage, the rotation angle of the object is estimated by a novel ring-shifted technique. The estimation is performed only for the most likely candidate which is the one having the highest similarity in the first stage. The experimental results show that the proposed method provides accurate estimation for template matching with arbitrary rotations and is applicable in various environmental conditions.

Due to global warming and climate changes, it is very important to use and conserve the power energy effectively. Monitoring the electrical consumption of consumers is one of the methods that can improve the energy usage efficiency. In this paper, a Non-Intrusive Load Monitoring (NILM) system, which applies a fuzzy classifier with the Fuzzy C-Means (FCM) clustering and optimization algorithms to identify the energizing and de-energizing statuses of each appliance, is proposed. Load energizing and de-energizing transient features are extracted, and the fuzzy classifier performs load identification based on these features. A two-stage fuzzy classifier is used in this paper. For the first stage, the FCM clustering is used to coarsely determine the parameters of the fuzzy classifier. Following this stage, two optimization algorithms, Error Back-Propagation Algorithm (EBPA) and Genetic Algorithm (GA), are employed to fine tune those parameters. As the classification results obtained from different realistic experimental environments, the proposed system is confirmed that it is able to identify the operational status of each appliance.