Showing papers by "Shun-Feng Su published in 2014"

Decomposed fuzzy systems and their application in direct adaptive fuzzy control.

Abstract: In this paper, a novel fuzzy structure termed as the decomposed fuzzy system (DFS) is proposed to act as the fuzzy approximator for adaptive fuzzy control systems. The proposed structure is to decompose each fuzzy variable into layers of fuzzy systems, and each layer is to characterize one traditional fuzzy set. Similar to forming fuzzy rules in traditional fuzzy systems, layers from different variables form the so-called component fuzzy systems. DFS is proposed to provide more adjustable parameters to facilitate possible adaptation in fuzzy rules, but without introducing a learning burden. It is because those component fuzzy systems are independent so that it can facilitate minimum distribution learning effects among component fuzzy systems. It can be seen from our experiments that even when the rule number increases, the learning time in terms of cycles is still almost constant. It can also be found that the function approximation capability and learning efficiency of the DFS are much better than that of the traditional fuzzy systems when employed in adaptive fuzzy control systems. Besides, in order to further reduce the computational burden, a simplified DFS is proposed in this paper to satisfy possible real time constraints required in many applications. From our simulation results, it can be seen that the simplified DFS can perform fairly with a more concise decomposition structure.

Moment Adaptive Fuzzy Control and Residue Compensation

Abstract: In this paper, a novel control scheme adopted from moment control is proposed. In the proposed approach, an adaptive fuzzy system is employed to learn the effective moment. It is easy to see that such an approach can avoid wild guessing for the effective moment, and as shown in our simulation, can have nice control performance. In traditional adaptive fuzzy control approaches, bounds of system functions are required to facilitate supervisory control so as to have the robust control property. It can be expected that when those bounds used in the supervisory controller are not proper, the output may not be able to follow the reference trajectory satisfactorily. With the proposed moment adaptive fuzzy control, the bound needed is only the supremum of the control variance between two consecutive steps. It is much easier to predict. In our study, in order to further relax this requirement, another adaptive system is employed to estimate the residue of the moment adaptive fuzzy control system. It is called residue compensation in this paper. It can be found that with residue compensation, the approach does not need a supervisory controller, but still can quickly track the reference in a satisfactory fashion. Various simulations are conducted to demonstrate the effectiveness of the proposed approaches.

Autonomous landing system for aerial mobile robot cooperation

Abstract: In vision-based autonomous landing, the accuracy of GPS signal and the efficiency of target tracking and detecting will affect the performance in autonomous landing system. This paper presents a method of cooperation among two aerial mobile robots at high altitude and low altitude for autonomous navigation and landing. Fully utilizing the extensive vision and high flexibility on the high-altitude aerial mobile robot to control the low-altitude aerial mobile robot to finish tracking and landing procedures. The flight controller achieves to track target and attitude control in real time. The flight controller using high level control which are fuzzy logic and neural network to give to low level control. The aerial mobile robots use the fuzzy logic control and neural network for positioning and maneuvering, and image-based visual servoing for tracking the low-altitude aerial mobile robot and landing target. By applying these methods, we have achieved a scalable and robust procedures for autonomous landing.

Multi-root I/O virtualization based redundant systems

Abstract: Redundancy, a method being designed to prevent failures due to software/hardware problem, is one of the most common applications in fault-tolerance systems. In this paper, we provide a multi-root I/O virtualization (MR-IOV) based redundant system architecture which supports high performance, reliability, and scalability to improve conventional redundant architecture with hardware multiplexer for the fail-over function. In order to fix this drawback, we proposed a redundant architecture to save these statuses in the shared memory, and the backup system will apply the states to fail-over primary host. From experiment results, we observe that the proposed architecture is feasible and it is better than the conventional redundant architecture.

Robust Gaussian Kernel Based Approach for Feature Selection

Abstract: The outlier problem of feature selection is rarely discussed in the most previous works. Moreover, there are no work has been reported in literature on symbolic interval feature selection in the supervised framework. In this paper, we will incorporate similarity margin concept and Gaussian kernel fuzzy rough sets to deal with the Symbolic Data Selection problem and it is also an optimizing problem. The advantage of this approach is it can easily introduce loss function and with robustness.

Fuzzy Based Compensation for Image Stabilization in a Camera Hand-Shake Emulation System

Abstract: The paper reports our study of image stabilization in a camera emulation system. In the study, a way of dealing with optical image stabilization is proposed to cope with blurring images caused by hand shake. The idea of the proposed image stabilization is to determine possible hand-shake situations based on the proposed rules and then to correct blurring image through position fuzzy control compensation. This method directly detects motion signals and employs a rule based mechanism to distinguish possible hand shake situations from normal camera movements. If a hand-shake situation is determined, the corresponding correction signal is generated to correct the image in a real-time fashion. In order to demonstrate the effectiveness of the proposed approach, in our implementation, the system directly moves the camera mounted on an X-Y platform to compensate the hand shake effects. Some statistics of the phenomena of the actual system are observed to form the scheme proposed in this study. From our experiments, the average percentage of shake reduction is 54.64%. It is clearly evident that this method indeed can have better image quality.

Dynamic obstacle avoidance path planning

Abstract: This paper proposes an image-based parallel lines distance measurement system (IBPLDMS), which is capable of executing an obstacle-detection and obstacle-avoidance path planning in a dynamic environment. The proposed IBPLDMS contains an image processing unit and an obstacle-avoidance unit. The image processing unit preprocesses captured images for an obstacle-avoidance task, in which the images are processed by the procedures of the grayscale transform, image morphologies, canny edge detection, connected-component labeling, and Hough transform. According to the definition of standard parallel lines, the obstacle-avoidance path for robots can be determined in the real time. Note that for the proposed IBPLDMS, only one webcam and the definition of the standard parallel lines are necessary for preparing for the setup. From the experimental results, we prove the proposed method is effective.

Feature genes selection and classification with SVM for microarray data of lung tissue

Abstract: Microarray analysis had been became a widely used tool for disease detection. It used tens of thousands of genes that would be a huge computational problem. The proposed approach applied feature genes selection and classification with support vector machine (SVM) for the microarray data of lung tissue. Based on the proposed approach, feature genes could be finding out according to the epsilon-support vector regression (epsilon-SVR) and selection ranked genes from each class. Moreover, applied multi-class support vector classification (multi-class SVC), cross-validation and parameter search methods to acquire great prediction classification accuracy and less computing time. That is, the effective dimension reduction for finding out feature genes is an important process in the proposed approach. The results shows the feature genes which our purposed approach finding out could acquire great prediction classification accuracy.

Two-loop PID control using PSO-RGA algorithm for solar heat pumps

Abstract: In this paper, a two-loop PID (proportional integral derivative) controller is presented for a class of solar water-source heat pump systems (SWSHPS), which provide warm water with desired temperature for a semiconductor factory. A combination of the well-known particle swarm optimization (PSO) and real-coded genetic algorithm (RGA) is proposed to off-line optimally find the controller parameters of the proposed control law. Through simulations, the two-loop PID control strategy is shown to be effective and robust against parameter variations. By comparing with two existing two-loop PID control strategies, the proposed two-loop PID controller, tuned by the new PSO-RGA algorithm, is shown to give superior transient performance, smaller steady-state errors and better robustness against parameter variations.

On Neural Fuzzy Systems

Abstract: Neural fuzzy system (NFS) is basically a fuzzy system that has been equipped with learning capability adapted from the learning idea used in neural networks. Due to their outstanding system modeling capability, NFS have been widely employed in various applications. In this article, we intend to discuss several ideas regarding the learning of NFS for modeling systems. The first issue discussed here is about structure learning techniques. Various ideas used in the literature are introduced and discussed. The second issue is about the use of recurrent networks in NFS to model dynamic systems. The discussion about the performance of such systems will be given. It can be found that such a delay feedback can only bring one order to the system not all possible order as claimed in the literature. Finally, the mechanisms and relative learning performance of with the use of the recursive least squares (RLS) algorithm are reported and discussed. The analyses will be on the effects of interactions among rules. Two kinds of systems are considered. They are the strict rules and generalized rules and have difference variances for membership functions. With those observations in our study, several suggestions regarding the use of the RLS algorithm in NFS are presented.

Direct adaptive control via decomposed fuzzy petri net

Abstract: This paper presents a novel direct adaptive controller design via decomposed fuzzy Petri net to solve the control tracking problem. The controller combines decomposed fuzzy system (DFS) and Petri net to achieve good performance with less computation time. In the DFS structure, fuzzy variables are decomposed into several layers. DFS has been shown to have fast learning capability but with a complicated system structure. In this study, Petri net is employed to form a mechanism in constructing meaningful component fuzzy systems in the DFS so that the number of fuzzy components can be dramatically reduced without significantly degrading the modeling performance. Finally, the effectiveness of the proposed controller scheme is verified by simulation results.

Study of terminal sliding mode control with fault-tolerant design

Abstract: This paper studies the terminal sliding mode control (TSMC) issue considering the fault-tolerant design. In the proposed scheme, TSMC can keep the merits of conventional sliding mode control (SMC), including robustness to uncertainties and/or disturbances, fast response, and easy implementation. Moreover, by using TSMC, the system states of TSMC will converge in finite time to the control objective point, i.e., equivalent point, after the systems states intersect the sliding surface. Simulation results show the advantages of the proposed method.

iPhone as Multi-CAM and Multi-viewer

Abstract: Recently, some of applications (APP) about the web of camera systems have been proposed for iPhone. The web of camera system is built on iOS smart mobile devices, and the objective-c programming language is employed to code applications in the Xcode. The iOS mobile devices are usually equipped with network and camera. Thus, it only needs to design software on the integration and links so as to replace the traditional webcam. The proposed system can be provides four iPhones or iPads to catching and watching the current images by the WiFi networks. In addition, the resolution of images and frame per second are also adjusted by the traffics of WiFi networks. In this study, the proposed system can be used as various applications such as home monitoring system and baby monitor system. The advantage is watch anytime, anywhere. And the mobile devices as mobile camera position can change location.

Robust interval support vector interval regression networks for interval-valued data with outliers

Abstract: Recently, the interval-valued data analysis is popular research topic in symbolic data analysis (SDA). For some of applications, it is natural to use interval-valued data because of uncertainty existing in the measurements, variability for defining a term (the minimum and the maximum temperature during a day), extremely behavior description (maximum wind speed in a given country), etc. However, the obtained data are always subject to outliers in some of applications. Moreover, the outliers may occur due to various reasons, such as erroneous measurements or noisy data from the tail of noise distribution functions. In order to handle the interval-valued data with outliers, a novel approach, called the robust interval support vector interval regression networks (RISVIRNs), is proposed. The RISVIRNs is extended from our previous work (e.g. the support vector interval regression networks; SVIRNs). It is easy to find that SVIRNs can have interval-valued data for outputs, but only can deal with the crisp input. Moreover, the outlier's effects are not discussed in SVIRNs. Hence, the support vector regression with interval-valued for input data (SVRI2) approach is proposed to determine the initial structure of RISVIRNs and to remove outliers form the interval-valued data set. Due to such approach can provide a better initial structure of RISVIRNs, the proposed approach can have fast convergent speed and robust against outliers. The experimental results with real data sets show the validity of the proposed RISVIRNs.