Current Issue

• 2017, No. 11　Published:25 November 2017 Last issue    Next issue
  Columns in this issue:

• Fast hardware implementation of on-chip

  Nor Flash wear leveling based on heap-sort

   

  XU Shu-tao1,HUANG Kai1,HUANG Kai-jie2,JIANG Xiao-wen1,ZHANG Xiao-meng1

  2017, 39(11): 1971-1979. doi:

  Abstract ( 145 )   PDF (1314KB) ( 258 )   　　 Review attachment

  Traditional implementation methods of Flash wear leveling mainly base on file system and focus on Nand Flash, while the wear leveling of Nor Flash is ignored. Nor Flash sometimes fails to be embedded in the operating system, and the cost can be too huge, so wear leveling cannot be implemented through the file system. We implement Flash wear leveling on hardware to solve this problem and reduce software cost. Four modules, which are wear leveling, address mapping, garbage collection and Flash interface unit, are implemented by Verilog. When a write request arrives, the sector which has the minimum erase time is found by the heap-sort, the virtual address is connected to the sector's physical address, and the address mapping list is updated. When the number of garbage sectors reaches a threshold value, garbage collection starts. Finally,  experimental results show that the operation time of initialization, heap deletion and read in hardware wear leveling algorithm is at most 14, 16.4 and 17.8 times faster than those of software algorithms respectively.

  An automated testing platform for high

  performance numerical simulation software

  TIAN Hong-yun1,LIU Qing-kai1,CHENG Jie2,YANG Zhang1,SHAN Ya-hui2

  2017, 39(11): 1980-1985. doi:

  Abstract ( 107 )   PDF (747KB) ( 242 )   　　 Review attachment

  To develop high performance numerical simulation software rapidly is the key to ensure the co-development of both high performance numerical simulation application and high performance computer, which is a bottleneck problem in the field of high performance computing application. Software testing is an indispensable part of software development, and it also has an important influence on software development efficiency. As a result, developing automated testing tools for numerical simulation software has great significance. As the existing automated testing platforms cannot meet the current demands well, we design and implement an auto-batch testing platform. Thanks to the unified description method with the test cases and loose coupling mechanism with the verification tools, our platform can be adapted to any numerical simulation software. With cross-platform nature and easy usability, users can check the test report through many ways, such as shell command line and web page, to fix the program's errors replied by the test logs. Our platform is an exploration and attempt for numerical simulation software testing. It has already provided testing and verification service for actual software development process. 19 projects and 1108 test cases from the CAEP Center have used our platform to do automatic acceptance tests. Compatible to a variety of testing targets and providing stable services during days of full system test, our platform demonstrates its general applicability and good stability.

  HEDSSA:a high efficient distributed

   stack for SIMD architecture

  SUN Hai-yan

  2017, 39(11): 1986-1990. doi:

  Abstract ( 98 )   PDF (550KB) ( 161 )   　　 Review attachment

  As the complexity and real-time requirement of applications become increasingly demanding, SIMD-vector processors with vector process element are widely used in various domains. The status of the programs running on processors is usually managed by compiler stack. The existing continuous compiler stack technique is not suitable for SIMD-vector processors. We propose a high efficient distributed stack for SIMD architecture called HEDSSA, which is more suitable for SIMD-vector architecture. Experimental results show that the HEDSSA can achieve higher performance in accessing data, calling functions, handling interrupt and dynamically allocating data.

   

  A fair resource allocation strategy based on

  preference in cloud environment

   

  HE Zhi-ming，LIU Min

  2017, 39(11): 1991-1999. doi:

  Abstract ( 111 )   PDF (1159KB) ( 178 )   　　 Review attachment

  Resource allocation strategies are an important research hotspot in cloud computing. How to consider the benefits of both cloud users and cloud providers, effectively meet the fairness requirement of system users and tasks, and make full use of system resources as much as possible are the main objectives. Considering the diverse user requirements in the cloud environment, each user requests a different number of tasks, and the resource requirements of each task are also different, we design a fair allocation strategy based on preference (FABP) and  define the concepts of user priority and task priority. Experimental results show that the algorithm can not only reduce the average task scheduling time, but also ensure the fairness of users and tasks and maximize the utilization of comprehensive resources.

   

  Universal assisted quantum computation

  ZHOU Xu,TAN Xiao-qing

  2017, 39(11): 2000-2005. doi:

  Abstract ( 114 )   PDF (537KB) ( 141 )   　　 Review attachment

  We devise a universal assisted quantum computation protocol. In this protocol the client Alice only has classical computers or limited quantum techniques, which is not sufficient for the universal quantum computation at her disposal. So Alice delegates her quantum computation to a remote quantum server Bob who is honest to execute the computation on his fully-fledged quantum computer. However, Bob learns nothing about Alice's input and output. Furthermore, our protocol only requires Alice to have the capacity of sending qubits and performing Pauli gates, with the properties of universality, half-blindness, correctness and verification.

  A max link selection scheme over independent

  and non-identically distributed fading channel

  JIA Xiang-dong1,2,XIE Man-gang1,ZHOU Meng1

  2017, 39(11): 2006-2015. doi:

  Abstract ( 93 )   PDF (2419KB) ( 138 )   　　 Review attachment

  We focus on the buffer-aided max link selection (MLS) relaying system in term of outage and delay performance over independent and non-identically distributed Nakagami-m fading channels. By using appropriate mathematical manipulation, the exact closed-form expressions of outage probability and average packet delay are achieved firstly, as well as the diversity and coding gains. The presented numerical results show that the non-identically distributed fading channels impose great loss in diversity gain for the MLS schemes. Moreover, the buffer-aided relaying systems are also unstable due to the fact that the average packet delays at all relays are different under this condition. Therefore, to improve the diversity and delay performance, we propose a novel weighted-based MLS (W-MLS) scheme. By employing weight vectors, the W-MLS scheme on one hand makes the equivalent powers of source-relay and relay-destination links equal. On the other hand, the W-MLS scheme also ensures that the equivalent power of relay-destination links is  greater than the one of source-relay links. As a result, comparing with the traditional MLS, the proposed W-MLS scheme can not only provide enough diversity and coding gains, but also enhance the stability of buffer-aided relaying systems due to that the average packet delays at all relays are entirely the same.

   

  A novel DOA estimation algorithm based on sparse

  representation under steering vector mismatch

  JIA Jin-hua，YU Jie-xiao，LIU Kai-hua，ZHAO Yu

  2017, 39(11): 2016-2021. doi:

  Abstract ( 102 )   PDF (646KB) ( 195 )   　　 Review attachment

  We present a new direction-of-arrival (DOA) algorithm which is applicable to the steering vector mismatch scenario based on sparse representation. In order to deal with the heavy-tailed noise in a real environment, we adopt the complex circularly symmetric generalized Gaussian distribution to simulate noise distribution. Given that the array steering vector can be transformed by sensor self motion and environmental factors, we employ the weighting least squares approach (WLS) to estimate the optimal gain value caused by  steering vector mismatch. Then we utilize sparse representation to reconstruct the signal model and convert the DOA estimation problem to a second-order cone programming (SOCP) problem. In order to reduce computational complexity, we adopt the singular value decomposition (SVD) method. Simulation results indicate that the proposed method can not only get high-resolution DOA estimation value, but  also effectively avoid the impact of steering vector mismatch on DOA estimation.

   

  A critical links detection method

  based on label propagation algorithm

  DONG Jian-liang,ZHAO Wen-tao,LI Fang,HUANG Xin-hao

  2017, 39(11): 2022-2027. doi:

  Abstract ( 128 )   PDF (760KB) ( 206 )   　　 Review attachment

  As network vulnerability gradually attracts people’s attention, detection of critical links (critical links problem) is playing an increasingly significant role in complex networks. According to the features of network society structure, and starting from the division of the network community, we are inspired by the Girvan-Newman (GN) algorithm and propose a critical links detection method on the basis of label propagation algorithm. Aiming at the resources wasting issue caused by assigning labels for each vertex in the algorithm iterative process, and the problem of instable results caused by random iteration, we bind the vertexes that have closer structure together by propagating labels once, and update the labels according to the sequence. Experiments show that the algorithm can figure out the critical links rapidly, stably and efficiently in a complex network.

   

  A novel multi-hop image transmission mechanism for

  energy balance in wireless multimedia sensor networks

  CHEN Xian-yi1,2,JIN Zhi-gang1,SU Yi-shan1

  2017, 39(11): 2028-2036. doi:

  Abstract ( 87 )   PDF (1051KB) ( 166 )   　　 Review attachment

  We propose a multi-hop image transmission (MHIT) mechanism for the heterogeneous wireless multimedia sensor network (WMSN) composed of common nodes and image nodes, where the images are compressed by the nodes in the nodes’ neighborhood. According to the hop count and distance, the image node determines whether or not the image should be compressed before transmission. If it costs more energy to compress the image, the image will be sent directly. Otherwise, the task of image compression is distributed to the nodes in the neighborhood, which can balance the network energy consumption and help ease the energy burden on the image nodes. Experimental results show that the proposed mechanism can effectively solve the problem of energy hole caused by the image compression in the WMSN, and significantly prolong the network lifetime. So it is well suited for remote image transmission in large-scale WMSN.

   

  Algebraic fault attack on elliptic curve

  scalar multiplication algorithms

   

  XU Sheng-wei1，CHEN Cheng1,2，WANG Rong-rong1,2

  2017, 39(11): 2037-2042. doi:

  Abstract ( 121 )   PDF (506KB) ( 149 )   　　 Review attachment

  Firstly, by analyzing the comb scalar multiplication and window non-adjacent form (NAF) scalar multiplication algorithms, we put forward an algebraic fault attack algorithm, which can recover all the private keys of elliptic curve cryptographic algorithms. The algebra fault attack algorithm cannot be detected in the execution process or fails when it meets all zero blocks. Then using the prime field curve of the commercial cryptography SM2 algorithm as the elliptic curve reference, the two scalar multiplication algorithms are attacked respectively in software simulation. It takes 13 minutes to attack the comb scalar multiplication algorithm and 18 minutes to the window NAF scalar multiplication algorithm, with 256-bits long private key recovered. In comparison, differential fault attack methods cannot attack the comb scalar multiplication algorithm, and they are vulnerable to "fault detection" and "zero block failure" simultaneously, so they fail. Experimental results show that the algebraic fault attack can be efficient to the scalar multiplications with precomputation, and the attacks are robust.

  A power amplifier model based on piecewise-linear

  functions and digital predistortion application

  JIA Bing,ZHAO Yu,LIU Kai-hua,MA Yong-tao,LIU Yan-bei

  2017, 39(11): 2043-2048. doi:

  Abstract ( 175 )   PDF (637KB) ( 270 )   　　 Review attachment

  Digital predistortion (DPD) becomes the most attractive technique in overcoming the nonlinearity and memory effects of power amplifiers (PAs) in communication systems. We propose a piecewise-linear memory polynomial (PWL-MP) model for DPD technique. Different from the generalized memory polynomial (GMP) model, our model converts the mulitplication of high order terms to the summation of PWL functions, in order to avoid the unstability of the muliplication of high order terms. As the threshold can be adjusted in the PWL-MP model, it is more flexible and accurate in the DPD technique. Experimental results show that the proposed model outperforms the GMP model while requiring less than 40% coefficients. The adjacent channel power ratio (ACPR) and the normalized mean square error (NMSE) can be improved by 1dB and 8 dB respectively. So the PWL-MP model can achieve a relatively better performance in the DPD technique.

   

  An improved ray tracing path searching algorithm

  YANG Jin-sheng，ZHAO Yue-qiu，QIU Guang-ran，CHEN Wei-gang

  2017, 39(11): 2049-2053. doi:

  Abstract ( 124 )   PDF (628KB) ( 253 )   　　 Review attachment

  The ray-tracing channel modeling based on the principle of geometrical optics can accurately predict multipath information, such as strength, time delay and angle of arrival and so on. The 3D models of experimental scenes are mostly segmented as polygons or triangles in this method. In order to improve the path searching efficiency of ray tracing in channel modeling, we propose an improved ray-triangle path searching algorithm. By adding the eliminating operations, the judging process is simplified, and the amount of computation is reduced. Taking three different types of scene model as example, algorithm comparison before and after the improvement is done by simulations. The results show that the improved path searching algorithm is faster than the original one. The more complex the scene model is, the higher the efficiency is.

   

  Abnormal pedestrian behavior recognition

   based on trajectory analysis

  HU Yuan1，XIA Li-min1，WANG Jia1,2

  2017, 39(11): 2054-2059. doi:

  Abstract ( 153 )   PDF (663KB) ( 816 )   　　 Review attachment

  We present a novel  abnormal behavior detection method  based on trajectory segment and topic model. In order to solve the problem of trajectory discontinuity caused by track deviation, all trajectories are firstly clustered by the fuzzy clustering algorithm, and then the sampling points of each segment of trajectory class are clustered by the latent Dirichlet allocation (LDA) topic model. The point of maximum probability is used as the visual word, and each trajectory class is represented by a series of visual words. In this sense, the local hidden Markov model (HMM) is established between every two visual words to detect abnormal trajectories through the  matching path method. Experimental results show that the proposed method can identify a variety of abnormal behavior, and improve the accuracy of abnormal behavior detection in CAVIAR database.

   

  Flame foreground extraction under

  the interference of intense light

   

   

  SU Xiang-ge,ZHANG Wei

  2017, 39(11): 2060-2065. doi:

  Abstract ( 106 )   PDF (669KB) ( 140 )   　　 Review attachment

  Flame foreground extraction is an important step in video-based fire detection algorithms, and it is the basis of following flame feature recognition algorithms. Given the fact that existing flame foreground extraction algorithms cannot extract flame foreground accurately under the interference of intense light or when the background color is similar to the flame, we propose a new algorithm to extract fire foreground. Firstly, the algorithm determines the primary suspected region by calculating the momentary motion region and the fire colored region. Secondly, we obtain the moving foreground region by different background subtraction methods according to the primary suspected region. Finally, we get the fire foreground region according to the moving foreground region and the fire colored region. Experimental results show that compared with four existing flame foreground extraction methods, the proposed algorithm has a much higher accuracy rate (100%). The proposed method can adapt to different complicated backgrounds and meet real-time requirement.

   

  A novel mixture spatial  Bayesian network model

  and its application in image segmentation

  CHEN Yuan-tao1,2，LIU Xuan-he1,2

  2017, 39(11): 2066-2073. doi:

  Abstract ( 88 )   PDF (1554KB) ( 183 )   　　 Review attachment

  Mixture proportions of the existing work don't have clear probability vector models, and some models cannot solve the iterative convergence problem of Markov chain Monte Carlo (MCMC). According to the Gaussian mixture models with spatial smoothing based on constraint, we present a new Bayesian model which can be applied in image segmentation. We use the probability density model of the latent Dirichlet allocation (LDA)  and the latent Dirichlet parameters of Gauss-Markov random field (MRF) to achieve parameter smoothing process. The proposed model has two advantages: 1) the model for the spatial smoothing constraint defines the probability vector model proportion; 2) we use the maximum a-posteriori (MAP) and the expectation maximization (EM) to achieve the update of closed parameters. Experimental results show that the proposed model has better image segmentation performance than the GMM method, and it has been successfully applied in the image segmentation of natural images and natural artistic images with noise.

  Binaryzation of real-valued curve descriptor

  WANG Zhi-heng,CHEN Lu-lu,LIU Hong-min,WANG Jing

  2017, 39(11): 2074-2085. doi:

  Abstract ( 102 )   PDF (2125KB) ( 177 )   　　 Review attachment

  Curve matching plays an important role in pattern recognition, computer vision and image understanding. With the widespread use of mobile devices, it is necessary to study the binary curve descriptors with small storage space and fast matching speed. We quantize the curve descriptors, such as MSCD, IOMSD, IOCD and TCHP, into binary strings represented by 0 and 1 by thresholding the real-valued vectors of the descriptors. Experimental results show that under the conditions of image rotation, viewpoint change and illumination change, the proposed binary descriptors yield a little higher matching accuracy while only requiring 1/32 or 1/16 storage space of the original curve descriptors.

   

   

  Visual analysis of Movielens data

  XU Bing-han,SHANG Hong-yun,MA Can,LI Shang

  2017, 39(11): 2086-2094. doi:

  Abstract ( 271 )   PDF (1246KB) ( 265 )   　　 Review attachment

  As we can collect increasingly growing movie data, study on movie data brings some inspiration. For example, analysis of the evolution of film genres can provide directors with recommendations on movie themes, and analysis of the relationship between economy and movie can find some reasons of film evolution. Study on high-score movies’ regularity over time can guide directors to choose movies’ release date. However, general research methods are inadequate to discover and present patterns hidden behind movie data visually because a film contains a movie name, genre information, ratings and other multiple attributes. We therefore adopt a visualization and visual analysis method for analyzing films and design a series of associated visualization views, which can analyze the time evolution of films in each genre from multiple temporal scales. We can also discover the correlation between movies and economy through the growth curve, and design a pie view to find patterns of high-score movies over time and across genre.

  A collaborative filtering algorithm based on time effect

  WU Fei,YU La-sheng,FENG Mei

  2017, 39(11): 2095-2101. doi:

  Abstract ( 113 )   PDF (681KB) ( 190 )   　　 Review attachment

  Collaborative filtering algorithms have been successfully applied in personalized recommendation. However, it is hard for traditional filtering algorithms to make sure that the accuracy and reliability of the nearest neighbors set is good enough because they ignore the impact of time factor. Even though there are a number of improved collaborative filtering algorithms, they cannot integrate time factor and user scores in their calculation. We propose a new algorithm based on time factor and our original work, which introduces time factor into user score prediction and user similarity calculation and assigns each item a dynamic weight by synthesizing time and user similarity. We finally obtain a Top-N set by filling the user-item matrix with user prediction scores and secondary calculating user similarity matrix. Experiments prove that the improved algorithm can enhance  recommendation accuracy and quality.

   

  A collaborative filtering recommendation algorithm

  based on familiarity similarity of cloud model

  WANG Jun,ZHU Jian-jun,QIN Lang

  2017, 39(11): 2102-2108. doi:

  Abstract ( 65 )   PDF (670KB) ( 181 )   　　 Review attachment

  In order to overcome the shortcomings of similarity measurement in recommendation algorithms, we propose a new collaborative filtering recommendation algorithm for calculating the integrated similarity of cloud model based on shape similarity and distance similarity. Taking account of the interest of users, we introduce the concept of familiarity similarity of cloud model, based on which the neighbors are identified, thus obtaining the recommendation results. Experimental results show that the proposed method can improve recommendation accuracy .

   

  A personalized anonymous publishing method in digital library

  JIA Jun-jie,CHEN Fei

  2017, 39(11): 2109-2114. doi:

  Abstract ( 74 )   PDF (482KB) ( 156 )   　　 Review attachment

  Aiming at the user privacy protection technology in released digital library data, we propose a personalized anonymous method. The user sets sensitive factors by themselves, and assigns attribute weights according to association rules among data attributes, thus obtaining the user information privacy protection degree of the data. We utilize this degree to conduct data division and anonymity, and realize personalized anonymous protection for the user. Experimental results show that the personalized parameters obtained through the weights of attributes can better reflect actual data relationship, decrease "excessive" protection caused by users' personalized settings, and in the meantime improve data publishing quality.