In this paper, we present a comparative study of four voting algorithms in two observer-based fault-tolerant control (FTC) architectures for an electric vehicle (EV) induction motor drive. The first architecture, call...
详细信息
In this paper, we present a comparative study of four voting algorithms in two observer-based fault-tolerant control (FTC) architectures for an electric vehicle (EV) induction motor drive. The first architecture, called output FTC, is based on the mechanical sensor, an EKF and a second-order sliding mode observer (SMO2). The second one, input FTC, is based on three controllers (PI, H loop shaping and the generalized internal model control), the most appropriate being selected to ensure good behaviour in presence of a multiplicative sensor fault (the fault is modelled as an exponential type emulating a bias). A third architecture, called hybrid FTC, based on the previous output and input fault-tolerant schemes, is built to mitigate simultaneous faults. Simulation and experimental results for a 7.5-kW induction motor drive show the efficiency of the approaches and their robustness against parametric variations for different load conditions.
This paper presents a novel method to simultaneously detect multiple trajectories of space debris in an observation image sequence to establish a reliable model for space debris environment in Geosynchronous Earth Orb...
详细信息
This paper presents a novel method to simultaneously detect multiple trajectories of space debris in an observation image sequence to establish a reliable model for space debris environment in Geosynchronous Earth Orbit (GEO). The debris in GEO often appear faintly in image sequences due to the high altitude. A simple but steady way to detect such faint debris is to decrease a threshold value of binarization applied to an image sequence during preprocessing. However, a low threshold value of binarization leads to extracting a large number of objects other than debris that become obstacles to detect debris trajectories. In order to detect debris from binarized image frames with massive obstacles, this work proposes a method that utilizes a cascade of numerical evaluations and a voting scheme to evaluate characteristics of the line segments obtained by connecting two image objects in different image frames, which are the candidates of debris trajectories. In the proposed method, the line segments corresponding to objects other than debris are filtered out using three types of characteristics, namely displacement, direction, and continuity. First, the displacement and direction of debris motion are evaluated to remove irrelevant trajectories. Then, the continuity of the remaining line segments is checked to find debris by counting the number of image objects appearing on or close to the line segments. Since checking the continuity can be regarded as a voting scheme, the proposed cascade algorithm can take advantage of the properties of voting method such as the Hough transform, i.e., the robustness against heavy noises and clutters, and ability of detecting multiple trajectories simultaneously. The experimental tests using real image sequences obtained in a past observation campaign demonstrate the effectiveness of the proposed method.
As one of the most challenging and attractive problems in the pattern recognition and machine intelligence field, imbalanced classification has received a large amount of research attention for many years. In binary c...
详细信息
As one of the most challenging and attractive problems in the pattern recognition and machine intelligence field, imbalanced classification has received a large amount of research attention for many years. In binary classification tasks, one class usually tends to be underrepresented when it consists of far fewer patterns than the other class, which results in undesirable classification results, especially for the minority class. Several techniques, including resampling, boosting and cost-sensitive methods have been proposed to alleviate this problem. Recently, some ensemble methods that focus on combining individual techniques to obtain better performance have been observed to present better classification performance on the minority class. In this paper, we propose a novel ensemble framework called Adaptive Ensemble Undersampling-Boost for imbalanced learning. Our proposal combines the Ensemble of Undersampling (EUS) technique, Real Adaboost, cost-sensitive weight modification, and adaptive boundary decision strategy to build a hybrid algorithm. The superiority of our method over other state-of-the-art ensemble methods is demonstrated by experiments on 18 real world data sets with various data distributions and different imbalance ratios. Given the experimental results and further analysis, our proposal is proven to be a promising alternative that can be applied to various imbalanced classification domains. (C) 2017 Elsevier Inc. All rights reserved.
A method for terrain classification based on vibration response resulted from wheel-terrain interaction is presented. Four types of terrains including sine,gravel,cement and pebble were *** vibration data were collect...
详细信息
A method for terrain classification based on vibration response resulted from wheel-terrain interaction is presented. Four types of terrains including sine,gravel,cement and pebble were *** vibration data were collected by two single axis accelerometers and a triaxial seat pad accelerometer,and five data sources were utilized. The feature vectors were obtained by combining features extracted from amplitude domain,frequency domain,and time-frequency domain. The ReliefF algorithm was used to evaluate the importance of attributes; accordingly,the optimal feature subsets were selected. Further,the predicted class was determined by fusion of outputs provided by five data sources. Finally,a voting algorithm,wherein a class with the most frequent occurrence is the predicted class,was employed. In addition,four different classifiers,namely support vector machine,k-nearest neighbors,Nave Bayes,and decision tree,were used to perform the classification and to test the proposed method. The results have shown that performances of all classifiers are ***,the proposed method is proved to be effective.
Various studies have examined the quality of one's sleep and further investigated several sleep disorders. In those investigations, accurately classifying one's sleep into the standardized sleep stages is impo...
详细信息
ISBN:
(纸本)9781538616451
Various studies have examined the quality of one's sleep and further investigated several sleep disorders. In those investigations, accurately classifying one's sleep into the standardized sleep stages is important. The conventional classification heavily depends on the manual examination of each expert on one's physiological signals during the sleep. Therefore, various automatic classification models have been proposed using the machine learning. Although they properly classify the sleep stages on average, there have been few investigations to specifically improve the classification accuracy of certain stages. Accurate determination of several stages considerably correlating with a disorder gives us a more effective hint to conquer the disorder. Accordingly, we propose a configured classification model focusing on the interesting sleep stages related to a challenging sleep disorder, the nocturnal enuresis. We consider the deterministic physiological signals of the interesting stages when training the classifiers. Further, the proposed system utilizes recurrent neural network to effectively learn the sequential feature of the physiological data. Our proposed system achieves the classification accuracy by 83.6% over the data. In particular, technique presents up to 15.5% higher accuracy to differentiate interesting stages than the support vector machine approach for the nocturnal enuresis.
There are numerous fault management techniques that can be used to enhance the reliability of a real-time critical system. Fault masking is one of the significant techniques amid those. It is one of the key approaches...
详细信息
ISBN:
(纸本)9781467311564
There are numerous fault management techniques that can be used to enhance the reliability of a real-time critical system. Fault masking is one of the significant techniques amid those. It is one of the key approaches to improve or maintain the normal behaviour of a system in the appearance of fault. voting as fault masking method involves the derivation of an output data object from a collection of n input data objects, as prescribed by the requirements and constraints of a voting algorithm. In data fusion, voting is a probable method of combining different data delivered by several sources (e.g. sensors) whose outputs may be mistaken. This paper proposes a new concept of classifying voting algorithms with some new parameters of evaluation. Then using this new concept and some widely known voting algorithms, we propose a hybrid history based weighted voting algorithm. We show that the proposed algorithm gives better and stable performance, in different error ranges, compared to the existing standard voting algorithms.
In this paper, we propose an experimental study for hybrid voting algorithm based on observer and robust control design for two mechanical sensor faults. The proposed strategy is applied for the induction motor speed ...
详细信息
ISBN:
(纸本)9781509012916
In this paper, we propose an experimental study for hybrid voting algorithm based on observer and robust control design for two mechanical sensor faults. The proposed strategy is applied for the induction motor speed drive of electrical vehicle powertrain. To adopt the best performance method for electrical vehicle application, we illustrate the effectiveness of the hybrid voting algorithm approach with the New European Driving Cycle (NEDC) speed profile. The experimental results demonstrate the effectiveness of the proposed Input/Output FTC architecture.
作者：
Yikun LiXiaoyuan DongSchool of Mathematics
Physics and Software Engineering Lanzhou Jiaotong University Mailbox 504 Lanzhou Jiaotong University Lanzhou 730070P.R.China
This paper proposes a semantic inference approach, which utilizes a group of context-sensitive Bayesian networks to infer the semantic concepts based on different regional spatial relationships, i.e. disjoined, border...
详细信息
This paper proposes a semantic inference approach, which utilizes a group of context-sensitive Bayesian networks to infer the semantic concepts based on different regional spatial relationships, i.e. disjoined, bordering, invaded by, surrounded by, near, far, right, left, above and below. Each Bayesian network performs the inference based on one kind of the regional spatial relationships. Finally, a voting algorithm is proposed to combine the group of the Bayesian networks into a more accurate and robust semantic concept classifier. The experiments using IKONOS imagery show that the precision of the proposed voting algorithm is consistently higher than that of the single context-sensitive Bayesian network.
This paper proposes a semantic inference approach, which utilizes a group of context-sensitive Bayesian networks to infer the semantic concepts based on different regional spatial relationships,***,bordering,invaded b...
详细信息
This paper proposes a semantic inference approach, which utilizes a group of context-sensitive Bayesian networks to infer the semantic concepts based on different regional spatial relationships,***,bordering,invaded by,surrounded by,near,far,right,left,above and *** Bayesian network performs the inference based on one kind of the regional spatial ***,a voting algorithm is proposed to combine the group of the Bayesian networks into a more accurate and robust semantic concept *** experiments using IKONOS imagery show that the precision of the proposed voting algorithm is consistently higher than that of the single context-sensitive Bayesian network.
暂无评论