AVIC Computing Technique Research Institute

Fengcheng, China

AVIC Computing Technique Research Institute

Fengcheng, China
Time filter
Source Type

Wen P.,China Institute of Technology | Li Y.,AVIC Computing Technique Research Institute | Zhang Y.,AVIC Computing Technique Research Institute
Proceedings - 12th International Conference on Computational Intelligence and Security, CIS 2016 | Year: 2016

The multiple unmanned aerial vehicles (UAVs) executing a same task coordinately based on formation flight have been more and more applied in the military and civil domains. To maximize the effectiveness of the multiple UAVs, a new and efficient control system is designed. After reasonably allocating the operator's and the UAVs' functions with a variable autonomous level, an airborne intelligent decision-making system dependent on the developed expert system is adopted to cope with the autonomous task environment uncertainty. In the aspect of the operator, the intelligent control methods like voice, gesture, eye tracking and brain computer interface are applied to realize the natural and efficient human-machine (operator-UAVs) interaction. Simulation experiments demonstrated the practicability and good performance of the designed control system. © 2016 IEEE.

Zhang J.-J.,Xidian University | Fang M.,Xidian University | Wang H.,Xidian University | Wang H.,AVIC Computing Technique Research Institute | Li X.,Xidian University
Engineering Applications of Artificial Intelligence | Year: 2015

High dimensionality of label space poses crucial challenge to efficient multi-label classification. Therefore, it is needed to reduce the dimensionality of label space. In this paper, we propose a new algorithm, called dependence maximization based label space reduction (DMLR), which maximizes the dependence between feature vectors and code vectors via Hilbert-Schmidt independence criterion while minimizing the encoding loss of labels. Two different kinds of instance kernel are discussed. The global kernel for DMLRG and the local kernel for DMLRL take global information and locality information into consideration respectively. Experimental results over six categorization problems validate the superiority of the proposed algorithm to state-of-art label space dimension reduction methods in improving performance at the cost of a very short time. © 2015 Elsevier Ltd. All rights reserved.

Zhao J.,AVIC Computing Technique Research Institute | An J.,Xi'an University of Science and Technology
International Journal of Pattern Recognition and Artificial Intelligence | Year: 2012

The contours and segments of objects in digital images have many important applications. Contour extractions of gray images can be converted into contour extractions of binary images. This paper presents a novel contour-extraction algorithm for binary images and provides a deduction theory for this algorithm. First, we discuss the method used to construct convex hulls of regions of objects. The contour of an object evolves from a convex polygon until the exact boundary is obtained. Second, the projection methods from lines to objects are studied, in which, a polygon iteration method is presented using linear projection. The result of the iteration is the contour of the object region. Lastly, addressing the problem that direct projections probably cannot find correct projection points, an effective discrete ray-projection method is presented. Comparisons with other contour deformation algorithms show that the algorithm in the present paper is very robust with respect to the shapes of the object regions. Numerical tests show that time consumption is primarily concentrated on convex hull computation, and the implementation efficiency of the program can satisfy the requirement of interactive operations. © World Scientific Publishing Company.

Li X.,Xidian University | Fang M.,Xidian University | Wang H.,Xidian University | Wang H.,AVIC Computing Technique Research Institute | Zhang J.-J.,Xidian University
Pattern Recognition Letters | Year: 2015

When there are a few labeled images, the classifier trained performs poorly even we use sparse coding technique to process image features. So we utilize other data from related domains as source data to help classification tasks. In this paper, we propose a Supervised Transfer Kernel Sparse Coding (STKSC) algorithm to construct discriminative sparse representations for cross domain image classification tasks. Specifically, we map source and target data into a high dimensional feature space by using kernel trick, hence capturing the nonlinear image features. In order to make the sparse representations robust to the domain mismatch, we incorporate the Maximum Mean Discrepancy (MMD) criterion into the objective function of kernel sparse coding. We also use label information to learn more discriminative sparse representations. Furthermore, we provide a unified framework to learn the dictionary and the discriminative sparse representations, which can be further used for classification. The experiment results validate that our method outperforms many state-of-art methods. © 2015 Elsevier B.V. All rights reserved.

Niu W.,AVIC Computing Technique Research Institute | Niu W.,Airborne | Niu W.,Xidian University | Cheng J.,Xidian University
AIAA Modeling and Simulation Technologies Conference, 2016 | Year: 2016

With basic requirements for fault diagnosis of health management technology—detect and diagnosis fault precisely, decrease CFAR efficiently, make a research from three respects: uncertainty of data, uncertainty of diagnosis result and uncertainty of feature parameter selection. Based on advantages for uncertainty problem of Bayesian network, Hierarchical Bayesian for fault diagnosis is proposed. The existing algorithms are not capable of selecting variables systematically so that they generally use the full model, which may contain unnecessary variables as well as necessary variables. Ignoring this model uncertainty often gives rise to, so called, the smearing effect in solutions. Complexity and difficulty of modeling is increased. The simulation results show this method can get better fault feature, improve fault discernment, and validate the model efficiency. © 2016 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.

Zhou M.,Huazhong University of Science and Technology | Li R.,Huazhong University of Science and Technology | Shang L.,Beihang University | Zhang L.,AVIC Computing Technique Research Institute
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) | Year: 2013

This paper proposes the design of a fault-tolerant controller area network (CAN). Several candidate redundancy strategies are introduced to improve the reliability. To choose a suitable one wisely, it is meaningful to make comparison on the reliability between the candidates. However, as there are redundant channels, the reliability of the network depends on the transmission requirements, which is difficult to evaluate. In this paper, a general approach based on minimal path set is proposed to model the reliability of complex fault-tolerant systems. With the proposed approach, the transmission-aware reliability of the redundant CAN network is evaluated. Comparisons on the reliability are carried out between the candidate strategies. In addition, comparisons on the cost, complexity and latency are also carried out. The results show the channel redundancy is superior to other strategies due to its high reliability, low latency and acceptable cost. © Springer-Verlag 2013.

Han W.,Dep. of Computer Science and Technology | Han W.,AVIC Computing Technique Research Institute | Han W.,Airborne | Bai X.,Dep. of Computer Science and Technology | And 2 more authors.
Procedia Engineering | Year: 2015

Typical aerospace embedded computers have many characters in common, especially in the aspect of the design of computer architecture. To evaluate this category of computers, 3 typical aerospace embedded computers were sampled and their characters of architecture and common technology were analyzed. Eight characters of architecture were concluded, and three evaluation dimensions were proposed. A basic guideline and a practical reference value or design were attached to each architecture character. The relationship was analyzed among the final assessment result, the evaluation dimensions and the architecture characters. Any aerospace embedded computer can be scored according to the assessment model for various estimate requirements. The evaluating method and progress were illustrated through one kind of advanced aerospace embedded computer's evaluation case. The case indicates that the assessment model is clear, effective and practical. It can provide some guidance for the design and performance analysis of aerospace embedded computers. © 2015 The Authors.

Li H.,Chang'an University | Wu Q.,Chang'an University | Dou A.,AVIC Computing Technique Research Institute
Journal of Information and Computational Science | Year: 2013

A novel algorithm is presented in this paper to detect anomalous vehicle behaviors such as abnormal stop and vehicle crashing. After objects detection, spatio-temporal trajectory of multiple objects can be obtained to construct the regional short-time constitute velocity model. Then gray model theory is adopted to estimate the motion model parameters. With the definition of abnormal traffic events patterns, individual vehicle movement can be judged as an abnormal or normal vehicle behavior. Experimental results show the proposed method can detect the abnormal traffic events effectively. © 2013 Binary Information Press.

Wen P.,AVIC Computing Technique Research Institute | Zhang Y.,AVIC Computing Technique Research Institute | Wang X.,Tianjin University | Wei H.,University of Reading
Hongwai yu Jiguang Gongcheng/Infrared and Laser Engineering | Year: 2013

A novel wedge-ring poor-pixels photoel ectric detector is valuable for a micro vision system. However an image acquired by the detector has extremely low resolution and it does not reflect the same or similar shape information of an object in the real world. To enable such a detector and its images available in further object identification, a unique shape recovery framework was presented in this paper. By rotating the wedge-ring detector around its center in a sub-wedge range, original low-resolution images were generated. Then linear interpolation along with a least squares method was applied to preliminarily recover the object shape. After noise removal via a two-stage level set evolution with an edge indicator function, the final high-quality object shape was achieved. Experiments demonstrate the effective performance of the proposed algorithms, in which the shape recovery rate is up to 95%.

Tang Q.,AVIC Computing Technique Research Institute | Feng X.D.,AVIC Computing Technique Research Institute
Applied Mechanics and Materials | Year: 2013

This paper mainly expounds the design and implementation of IEEE 1394 signaling rate converter, and emphatically describes the system structure design, basic hardware design, and function realization of IEEE 1394 signaling rate converter. IEEE 1394 signaling rate converter mainly realizes the function of signaling rate conversion in legacy mode, signaling rate conversion in Beta mode, mode change between legacy and Beta, and 1394 bus relay. © 2013 Trans Tech Publications Ltd, Switzerland.

Loading AVIC Computing Technique Research Institute collaborators
Loading AVIC Computing Technique Research Institute collaborators