Table of Content

30 September 2015, Volume 33 Issue 5

Previous Issue    Next Issue

Communication Engineering

Subspace Tracking-Based Semi-blind Channel Estimation for Massive MIMO Systems

XU Feng-yang, WANG Dong, XIAO Yang, KOU Jin-feng

2015, 33(5):  459-469.  doi:10.3969/j.issn.0255-8297.2015.05.001

Asbtract ( 1130 )   PDF (5506KB) ( 818 )  

References | Related Articles | Metrics

Semi-blind channel estimation based on eigenvalue decomposition (EVD) or singular value decomposition (SVD) requires repeated computation of EVD or SVD, both having high computational complexity. Such method is unsuitable for practical implementations in multicell multiuser massive MIMO (MU-massive-MIMO) systems. A fast implementation of the semi-blind channel estimation scheme for multicell MU-massive-MIMO systems is presented in this paper. This scheme provides a new recursive method to compute the ambiguity matrix, mainly using the least-squrares (LS) and linear minimum mean-square error (LMMSE) principles. A fast recursive row-Householder (FRRH) subspace tracking algorithm is used to speed up estimation of the signal subspace of received vectors. Numerical results show that the proposed scheme has good estimation performance and can effectively mitigate the effect of pilot contamination.

A Novel Method for Generation of Correlated Composite Fading

ZHOU Sheng-kui, ZHU Qiu-ming, DAI Xiu-chao, LIU Xing-lin, CHEN Xue-qiang

2015, 33(5):  470-480.  doi:10.3969/j.issn.0255-8297.2015.05.002

Asbtract ( 715 )   PDF (1872KB) ( 665 )  

References | Related Articles | Metrics

The traditional channel models usually consider the effects of multipath fading only. A new fading model including shadowing and multipath fading is established, and a novel generation method for this correlated composite fading is presented. The proposed method is based on the sum-of-sinusoids (SoS) theory that is easy to implement with hardware. Moreover, theoretical envelope distribution and correlation coefficient of output fading are derived. A method for calculating simulation parameters is given. Experiments on a Virtex4 FPGA hardware platform show that the envelope distribution, Doppler power spectrum and correlation characteristic agree well with the expected values, meaning that the proposed method can be used to simulate random distortion of wireless signals caused by real propagation environment in the lab.

Differential Fault Analysis on Encryption Mode of MD5

SHEN Xuan, ZHAO Guang-yao, LI Chao, LI Rui-lin

2015, 33(5):  481-490.  doi:10.3969/j.issn.0255-8297.2015.05.003

Asbtract ( 1052 )   PDF (2212KB) ( 472 )  

References | Related Articles | Metrics

By studying differential properties of the round functions in the encryption mode of MD5, we propose a fast algorithm to solve the differential equation. We give a differential fault analysis on the encryption mode of MD5 from the third last round with the proposed algorithm. The result shows that, if we induce faults from the third last round, only 56 random faults in average are required to obtain 512 bit key successfully.

Complementary-Weighted Iterative Fusion Algorithm for Heterogeneous Dual 9-Axis MEMS Inertial Sensor Data

WANG Wei, ZHANG Jin-yi, ZHANG Hong-hui, CAI Chun-yan, LI Jian-yu

2015, 33(5):  491-501.  doi:10.3969/j.issn.0255-8297.2015.05.004

Asbtract ( 1217 )   PDF (10038KB) ( 987 )  

References | Related Articles | Metrics

In micro-electro-mechanical system (MEMS) inertial sensor navigation, traditional techniques only use single multi-axis sensor to navigate and get position of the moving target. However, measurements of sensors and noise characteristics are affected by environment conditions. Besides, acceleration error, gyro drift, platform angle error and the error of platform installation also have important influences. From the view point of data fusion of complementary non-single multi-axis sensor, this paper uses two 9-axis MEMS inertial sensors to collect the original information of moving targets, and develops a complementary-weighted iterative fusion algorithm. The raw data of heterogeneous dual 9-axis MEMS inertial sensor are preprocessed. Then, using minimum variance, the observed values are estimated with Kalman filtering. A model of weight updating is established to improve accuracy according to evaluation of variance and an innovation equation. Experimental results show that, comparing with the traditional method, the proposed algorithm can improve navigation accuracy by more than 50%.

Signal and Information Processing

Multi-feature Selection Tracking Based on Support Vector Machine

HU Zhao-hua, XU Yu-wei, ZHAO Xiao-lei, HE Jun, ZHOU You

2015, 33(5):  502-517.  doi:10.3969/j.issn.0255-8297.2015.05.005

Asbtract ( 1002 )   PDF (26802KB) ( 612 )  

References | Related Articles | Metrics

Discriminative tracking is generally based on a single feature and uses the current track result (a positive sample) and some negative samples to train the classifier. It may lead to tracking due to occlusion, illumination changes and deformation. To overcome the problem caused by single feature description and single positive sample training, we propose a multi-feature selection algorithm based on the support vector machine (SVM). The classifier is trained with multiple positive and negative samples based on multi-feature descriptions. In the tracking step, a candidate sample with maximum confidence probability is used as the tracking result. Further, we take advantage of a subspace learning method to update positive samples used in the classifier training. Experimental results show that the proposed algorithm has desirable performance in a variety of challenging situations.

Polarization Estimation of Received Interference Signal Based on Spatial Modulation Effect

LIU Wen-zhao, DAI Huan-yao, HUANG Zhen-yu, CUI Jian-ling

2015, 33(5):  518-526.  doi:10.3969/j.issn.0255-8297.2015.05.006

Asbtract ( 860 )   PDF (5517KB) ( 652 )  

References | Related Articles | Metrics

Precision of interference polarization estimation is a major factor influencing radar polarization anti-jamming capability. To a certain extent, the spatial polarization characteristic (SPC) of radar is ill-conditioned, resulting in unstable solution to the estimation of received wave polarization. A least squares method with a quadratic constraint is presented to estimate polarization of the received wave. It treats the ill-conditioning problem in estimating the polarization state based on least square methods. Factors affecting stability of polarization estimation are analyzed, with computer simulation results in ill-conditioned cases for particular antenna forms. Theoretical analysis is then made and algorithmic steps given according to the principle of the proposed method. The simulation shows that the method can significantly reduce variance of polarization estimation so as to improve the estimation accuracy, useful in improving radar polarization anti-jamming capability.

Joint DOD and DOA Estimation with High Accuracy in Bistatic MIMO Radar Using Sparse Array

CUI Chen, LIANG Hao, YU Jian

2015, 33(5):  527-540.  doi:10.3969/j.issn.0255-8297.2015.05.007

Asbtract ( 1139 )   PDF (6546KB) ( 887 )  

References | Related Articles | Metrics

Based on extended aperture of bistatic MIMO radar using asparse array, a joint high accurate method for estimating direction of arrival (DOA) and direction of departure (DOD) is proposed. A bistatic MIMO radar with sparse linear distributed subarrays and dual baselines for both transmission and receiving is used. By designing a nonuniform linear array and using Khatri-Rao product processing, the long baseline of inter-subarray spacing and short baseline of inter-sensor spacing in the transmission/receiving arrays are virtually extended simultaneously. With the redundancy removal and spatial smoothing for the received data, high accuracy estimation of DOA and DOD are accomplished using a dual-size unitary ESPRIT algorithm for angular ambiguities resolution. Compared with conventional algorithms, the proposed method can achieve virtual extension of aperture for the MIMO radar without increasing the number of sensors and computational complexity while achieving effective estimation for transmitting and receiving angle with better performance. Simulation results show effectiveness of the proposed method.

Tensor Voting Based Pavement Crack Extraction

LI Ai-xia, GUAN Hai-yan, ZHONG Liang, YU Yong-tao

2015, 33(5):  541-549.  doi:10.3969/j.issn.0255-8297.2015.05.008

Asbtract ( 1134 )   PDF (15487KB) ( 403 )  

References | Related Articles | Metrics

This paper proposes a multi-scale tensor voting framework that applies tensor voting to mobile laser scanning data to extract pavement cracks. Trajectory data are used to extract road curbs from profiles along the travelling line to separate road points from non-road points. The extracted road points are interpolated into road feature images. Thus curvilinear cracks are enhanced and extracted with a multi-scale tensor voting framework. Experiments on mobile laser scanning data and road image data were carried out. The results show that the method is robust to noise in both road images and feature images, and can achieve good performance in pavement crack extraction.

A Locality Sensitive Hashing Algorithm Based on CUDA

ZHANG Yi-fan, YU Xiao-qing, AN Xuan-dong, WAN Wang-gen

2015, 33(5):  550-558.  doi:10.3969/j.issn.0255-8297.2015.05.009

Asbtract ( 1203 )   PDF (6971KB) ( 706 )  

References | Related Articles | Metrics

The traditional locality sensitive hashing (LSH) algorithm often takes a large memory space and long settling time to build a hash table. In the query phase, it takes more than 95% of the overall running time to search K nearest neighbor data items of samples. To solve these problems, this paper uses the compute unified device architecture (CUDA) to transplant LSH algorithm into a GPU, using parallel multi-threads to calculate the values of data items to build the hash table. In the query phase, we introduce multisample query based on work queue to global memory to improve efficiency of the algorithm. Experimental results show that computing speed of the proposed LSH algorithm is 12 times faster than the traditional LSH algorithm.

Application of Building Information Extraction Based on High Resolution Optical and SAR Images

ZHANG Bin, ZHANG Yao-ming, ZHANG Zhi, QIN Qian-qing, WANG Han

2015, 33(5):  559-567.  doi:10.3969/j.issn.0255-8297.2015.05.010

Asbtract ( 1125 )   PDF (3521KB) ( 1093 )  

References | Related Articles | Metrics

Three-dimensional reconstruction of buildings based on high-resolution spaceborne remote sensing data is an important topic, with theoretical research and application values. Optical and SAR images are two types of widely applied remote sensing data. Highresolution remote sensing is an important trend in contemporary development. Building detection and height estimation are described and summarized. Physical scattering characteristics of buildings in high resolution SAR images are analyzed, and simulation carried out. Research status of building detection and height estimation is presented in the order of used image sources. In-depth analysis of the problems in building detection and height estimation is presented. Finally, the future research directions for building three-dimensional reconstruction based on high-resolution optical and SAR images are suggested.

Electronic Engineering

Strain Effect on Electrical Conductivity of Three-Dimensional Graphene Composites

ZHENG Chen-fei, XU Rong-qing, CHEN Jing, LU Yun-qing

2015, 33(5):  568-574.  doi:10.3969/j.issn.0255-8297.2015.05.011

Asbtract ( 831 )   PDF (14815KB) ( 701 )  

References | Related Articles | Metrics

A three-dimensional graphene foam and poly dimethyl siloxane (GF/PDMS) composite was fabricated by infiltrating PDMS into 3D GF. It was synthesized by chemical vapor deposition (CVD) with nickel foam as a template. Research of the strain effect on electrical conductivity of three-dimensional grapheme composites indicates that the relative change rate of resistance of the composites increases with the increaseof bending curvature under bending stress. The relative change rate increases quickly with a small bending curvature while slowly with a big curvature. Under tensile strain, the relative change rate of resistance of the composites increases with the increaseof tensile stress. Average strain sensitivity is about 6. The above results show that the GF/PDMS composite is potentially applicable to flexible conductors and stress sensing materials.