Loading...
Office

Table of Content

    20 August 2018 Volume 45 Issue 4
      
    Massive MIMO communication method based on the zero pilot in LOS environment
    LI Wengang;HU Aili;SUN Longxin;WANG Yiwei
    Journal of Xidian University. 2018, 45(4):  1-5+85.  doi:10.3969/j.issn.1001-2400.2018.04.001
    Abstract ( 472 )   PDF (558KB) ( 272 )   Save
    References | Related Articles | Metrics

    To resolve the problem that the pilot overheads and the computation complexity of traditional channel estimation algorithms are too large in frequency division duplex massive multiple input multiple output (FDD Massive MIMO) system, a massive MIMO communication method based on the zero pilot is proposed. The transmitter does not transmit the pilot. The transmitter uses the location information to calculate the direction of the arrival (DOA) angle, and then adjusts the transmitted signal to form the directional beam. The receiver directly demodulates the received data without channel estimation and equalization. Simulation results show that, compared with the traditional FDD Massive MIMO system, the proposed method greatly reduces pilot overheads and the computational complexity, and improves the spectral efficiency, and that the system sum-rate improves by about 4.6dB at a high signal-to-noise ratio.

    Improved approach to enhancing the robustness and effectiveness of networks
    MU Caihong;CHAI Wenyi;LIU Yi;LIU Jing
    Journal of Xidian University. 2018, 45(4):  6-11.  doi:10.3969/j.issn.1001-2400.2018.04.002
    Abstract ( 460 )   PDF (830KB) ( 161 )   Save
    References | Related Articles | Metrics

    Focusing on the problem of the vulnerability for scale-free networks under malicious attacks, an improved approach is proposed to enhance the robustness and effectiveness of the networks. Based on the existing robustness index and effectiveness index, this paper presents a new object function by combining these two indexes together reasonably. Then taking the onion-like structure as the heuristic information, a heuristic mixed searching operator is constructed. By adopting the simulated annealing algorithm combined with the heuristic mixed searching operator and a novel evaluation model based on the rate of change, the network structure is optimized iteratively, resulting in a network with higher robustness and effectiveness. Experimental results show that this new approach can improve the robustness and effectiveness of the initial network simultaneously.

    OFDM radar and communication sharing signal processing using the correlation function
    GU Yabin;ZHANG Linrang;ZHOU Yu;ZHAO Yonghong
    Journal of Xidian University. 2018, 45(4):  12-17+105.  doi:10.3969/j.issn.1001-2400.2018.04.003
    Abstract ( 388 )   PDF (931KB) ( 200 )   Save
    References | Related Articles | Metrics

    A signal processing method based on the correlation function is proposed for integration of radar and communication with OFDM. First, the Doppler shift between pulses is compensated, then the cross ambiguity function is taken between the compensated signal and transmitted sequence, and finally, the square of the modulus of the cross-ambiguity function output is regarded as the cost function. The range and velocity can be estimated from the peaks of the cost function in the Range-Doppler plane. The randomness property of the communication term has a significant influence on performance. In order to analyze the performance of the proposed method, the mathematical expression for the peak shape and sidelobe mean level is derived, and then the influence of Doppler on the performance is analyzed. Finally, simulation indicates the correctness of the proposed method.

    Estimation of parameters of Pareto distribution using Mellin transform
    CHEN Shichao;LUO Feng;LUO Meiyixiang;HU Chong
    Journal of Xidian University. 2018, 45(4):  18-22+117.  doi:10.3969/j.issn.1001-2400.2018.04.004
    Abstract ( 376 )   PDF (976KB) ( 141 )   Save
    References | Related Articles | Metrics

    Some traditional parameters estimation methods have some problems that the estimation range is limited, that the computation is complicated, and that the number of pulses accumulated is limited. In order to solve these problems, a novel parameter estimation method of sea clutter under Pareto distribution is proposed in this paper. In the proposed method, the probability density function is firstly transformed with Mellin transform. Then the second kind of characteristic function of Pareto distribution is computed under the different numbers of pulses, and the first two logarithmic cumulants of Pareto distribution are deduced in detail. Finally, a comparison between the proposed method and some traditional methods is made based on the Monte Carlo simulation experiments. Experimental results show that the proposed method not only can estimate the parameters in all defined domains with a high fitting precision, but also has a more concise analytical expression with a low computational complexity.

    Analyzing pan-cancer DNA methylation patterns via clustering
    YANG Liying;YANG Shengnan;YUAN Xiguo;GENG Fangge;ZHANG Junying
    Journal of Xidian University. 2018, 45(4):  23-28.  doi:10.3969/j.issn.1001-2400.2018.04.005
    Abstract ( 588 )   PDF (2100KB) ( 102 )   Save
    References | Related Articles | Metrics

    There have been many studies on DNA methylation, but most researches are for a single cancer, individual genes or smaller regions of the gene. In view of the problem, this paper proposes a clustering-based method and analyzes DNA methylation from the perspective of the Pan-cancer at the whole genome level. First, methylation levels of the multiple cancer types are analyzed by SAM and the differential methylation sites are screened out. Also, by calculating the correlation between methylation and gene expression, common regulatory sites are identified. Then AP clustering is carried out at differential methylation sites. Finally, GO and KEGG are adopted for gene annotation and enrichment analysis. Experiments are performed on six categories of cancers from the Pan-cancer project in TCGA. 2184 differential methylation sites and 9 clusters are obtained based on SAM and AP. Experimental results show that the relationship between methylation and gene expression is complex rather than simple positive or negative correlation. From the results of GO and KEGG, we also conclude that the corresponding genes in clusters have been enriched in multiple cancer-related pathways, and are of good biological interpretation.

    Blind channel estimation for MIMO-OFDM systems via  non-redundant block precoding
    YANG Ruonan;LOU Shuntian;ZHANG Weitao
    Journal of Xidian University. 2018, 45(4):  29-33.  doi:10.3969/j.issn.1001-2400.2018.04.006
    Abstract ( 348 )   PDF (504KB) ( 93 )   Save
    References | Related Articles | Metrics

    The conventional blind channel estimation algorithms for MIMO-OFDM systems use block precoding to avoid the imperfection of subspace methods, which causes the distortion of a part of transmit information. To resolve this problem, a new non-redundant block precoding algorithm is proposed. Transmit signals are reshaped by parity, and then precoded by a block precoder, so that the autocorrelation matrix of receive signals has a block structure. The two diagonal blocks are chosen to obtain channel estimation, respectively. Then conditions the precoder should meet and the impact on the channel estimation and system bits error produced by the precoder are discussed. Based on the discussion, a designed balance operation is applied in the two estimations to obtain the final estimation. Simulation results demonstrate that the block precoding method can lead to a better performance of blind channel estimation while achieving a lower bits error.

    Novel method of destroyed document reconstruction
    XING Nan;ZHANG Jianqi;LIU Pengfei;CAO Furong
    Journal of Xidian University. 2018, 45(4):  34-39.  doi:10.3969/j.issn.1001-2400.2018.04.007
    Abstract ( 299 )   PDF (1468KB) ( 81 )   Save
    References | Related Articles | Metrics

    To solve the problem that it is difficult to recover the paper document destroyed by a shredder, a novel method for destroyed document reconstruction is proposed. First, based on the structural characteristics of Chinese words, the structural block model of the character in shred is built, and the structural blocks are identified by a classifier. Second, the matching degrees between shreds are measured by the matching probability and the number of mismatches between structural blocks. Third, the two matching measures are fused to generate an evaluation function for the global matching of shreds. Finally, based on the evaluation function, the best matching of shreds is realized by the genetic algorithm. Experimental results show that the proposed method can effectively restrain the effect of information loss, etc.  on the document reconstruction, and that it achieves a higher reconstruction accuracy than the existing methods.

    Detecting water bridge in SAR images via a scene semantic algorithm
    HUANG Yong;LIU Fang
    Journal of Xidian University. 2018, 45(4):  40-44.  doi:10.3969/j.issn.1001-2400.2018.04.008
    Abstract ( 336 )   PDF (2451KB) ( 143 )   Save
    References | Related Articles | Metrics

    Much speckle noise causes the bridge to be missing or undetected in SAR images. Aiming at the problem, we propose a novel algorithm based on scene semantic water bridge recognition. First, the image is automatically segmented into land and water scenes which narrow the search region of the bridge to the junction area. Second, the Primal Sketch features of the image are extracted to suppress the interference of speckle noise. Last, the membership function of the bridge is defined by its geometric characteristics based on the Prime Sketch features. Experiments show that the proposed algorithm effectively reduces the missing rate and error detection rate of the bridge, and has better robustness.

    Capacitor selection method based on the effective decoupling rise time
    LIU Yang;XIA Jianqiang;CHU Xiuqin
    Journal of Xidian University. 2018, 45(4):  45-50+56.  doi:10.3969/j.issn.1001-2400.2018.04.009
    Abstract ( 366 )   PDF (762KB) ( 68 )   Save
    References | Related Articles | Metrics

    The traditional power distribution network design method is based on the frequency-domain target impedance, but this method ignores the transient characteristics of the power distribution network and current excitation, and can not represent the high frequency characteristics of the power distribution network accurately. Aiming at this problem, under the condition that the input current is in the triangular form, the critical rise time of the input current which needs to add decoupling capacitors and the effective decoupling rise time range of the capacitor are deduced. The relationship between the maximum voltage noise of the decoupling capacitor and the rise time and the relevant capacitance parameters is further investigated, and then the capacitor selection method based on the effective decoupling rise time is proposed. Compared with the frequency-domain decoupling method, the results of the decoupling design for 4 typical stimuli show that the number of capacitors can be reduced more than 31.6%.

    New construction methods for multipermutation codes under the Chebyshev distance metric
    ZHAO Peng;MU Jianjun;JIAO Xiaopeng
    Journal of Xidian University. 2018, 45(4):  51-56.  doi:10.3969/j.issn.1001-2400.2018.04.010
    Abstract ( 276 )   PDF (458KB) ( 73 )   Save
    References | Related Articles | Metrics

    For the limited-magnitude errors occurring in a flash memory system and the disadvantage that the information rates of rank-modulation permutation codes are not high enough, a simple construction method for regular multipermutation codes that can correct these errors is proposed by using the direct product operation on multipermutation sets which have the same minimum distance under the Chebyshev distance metric. Moreover, by using the projection operation on the known simple regular multipermutation codes, the multipermutation sets with different minimum distances are obtained. Based on the direct product of these multipermutation sets, a generalized construction method for regular multipermutation codes is proposed. Compared with the existing construction methods of multipermutation codes under the Chebyshev distance metric, the proposed construction methods of multipermutation codes based on direct product operation are more intuitive. Some calculation examples are provided to verify the validity of the proposed construction methods.

    Noise-robust multi-feature joint learning HRRP recognition method
    LI Long;LIU Zheng
    Journal of Xidian University. 2018, 45(4):  57-62.  doi:10.3969/j.issn.1001-2400.2018.04.011
    Abstract ( 331 )   PDF (749KB) ( 96 )   Save
    References | Related Articles | Metrics

    In order to improve the recognition performance under low signal-to-noise ratio (SNR) conditions, a novel method is proposed for radar target high range resolution profiles (HRRP). This method achieves good recognition performance based on multi-feature joint learning for noisy HRRPs. The framework of this method is constructed based on sparse representation and low-rank representation, which are applied to extract the local and global features of target HRRPs. In the training stage, a feature extraction dictionary is produced based on the joint learning structured analytical discriminative dictionary method to improve the recognition performance. The cancellation method is implemented for noise suppression in the testing stage. Experimental results on the measured HRRP data demonstrate that the proposed method can significantly improve the overall recognition performance for HRRP testing samples under relatively low SNR conditions with a satisfactory real-time ability.

    Dynamic virtual optical network mapping algorithm with cooperation between nodes and links
    YANG Fan;QIN Shuai;XU Zhanqi;MA Tao;DING Zhe
    Journal of Xidian University. 2018, 45(4):  63-68.  doi:10.3969/j.issn.1001-2400.2018.04.012
    Abstract ( 417 )   PDF (554KB) ( 89 )   Save
    References | Related Articles | Metrics

    A dynamic virtual optical network mapping algorithm based on nodes and links cooperation is proposed to make multiple dynamic virtual optical networks which are logically isolated to efficiently share the underlying elastic optical network resources. The computational capacity of the node, the spectrum resource and the relationship between the nodes to be mapped and the nodes mapped are considered during node mapping in this algorithm. When virtual nodes are mapped into the physical nodes, link mapping is considered. The virtual nodes around a virtual node are mapped into the nodes around a physical node. The method uses the cooperation between nodes and links to avoid long hop physical routes during link mapping. Simulation results show that the performance of the blocking ratio and link utilization of the algorithm are significantly improved. This algorithm can save the network resource effectively.

    Analysis of echo correlation in diversity MIMO radars
    LIU Jieyi;ZHANG Linrang;YU Hengli;TANG Shiyang;ZHAO Shanshan
    Journal of Xidian University. 2018, 45(4):  69-74.  doi:10.3969/j.issn.1001-2400.2018.04.013
    Abstract ( 393 )   PDF (753KB) ( 79 )   Save
    References | Related Articles | Metrics

    For the target echo correlation, the existing target model has a huge limitation which just represents the angle or frequency interval with respect to the station, but ignores the influence of target attitude. To solve this problem, this paper utilizes the four straight prism scattering model to better simulate the true target, and the research shows that the correlation coefficient is affected by the observed direction, target size, spatial and frequency interval at the same time in the diversity MIMO radar. The study focuses on the variation of the targets effective size with the observed direction and spatial size in the case that the radars stay at different positions, which influences the correlation coefficient severely.

    Clustering by samples movement in the superposition information entropy field
    XU Tuo;LI Jie;WANG Ying
    Journal of Xidian University. 2018, 45(4):  75-79+154.  doi:10.3969/j.issn.1001-2400.2018.04.014
    Abstract ( 382 )   PDF (2389KB) ( 139 )   Save
    References | Related Articles | Metrics

    The essence of clustering is to compare the similarities among samples on different scales. Due to the complexity of the spatial distribution of sample data, the similarity measurement in the process of repeatability, and the algorithm for adaptive problems, the clustering algorithm cannot lead to the correct result in the process of data clustering. In order to solve the complex problem of spatial distribution of sample data, we present a data migration clustering algorithm based on superposition information entropy, which is used to construct the entropy field of the data in the numerical space, and the datawandering to implement data segmentation and complete clustering. Experimental results show that this method can not only obtain a better clustering effect, but also have data adaptability.

    Text image refocusing by using the convolutional neural network
    WANG Kangkang;WANG Keyan;LI Yunsong
    Journal of Xidian University. 2018, 45(4):  80-85.  doi:10.3969/j.issn.1001-2400.2018.04.015
    Abstract ( 288 )   PDF (2165KB) ( 106 )   Save
    References | Related Articles | Metrics

    We propose a new image refocusing algorithm based on the convolutional neural network. We analyze the traditional wiener filtering method, and derive it. We transform the frequency domain division into a cyclic convolution and decompose the kernel by SVD. After that, we design a new structure of the convolutional neural network by cyclic convolution and one-dimensional convolution. This network can not only simulate the defocusing process of wiener filtering, but also restore the image without a kernel, and have good anti-noise performance. At the same time, the convolutional neural network is fast-convergent and parameter insensitive.

    Scheme for generation of a highly adaptive physical layer secret key
    SHEN Zhiwei;LIU Jingmei;HAN Qingqing
    Journal of Xidian University. 2018, 45(4):  86-91.  doi:10.3969/j.issn.1001-2400.2018.04.016
    Abstract ( 373 )   PDF (777KB) ( 105 )   Save
    References | Related Articles | Metrics

    To enhance the performance of the secret key-based mechanism of physical layer security, a scheme for generation of a highly adaptive key using the orthogonal frequency division multiplexing (OFDM) systems channel response is proposed. First, linear minimum mean square error (LMMSE) estimation is adopted to improve the agreement of the secret key. Second, a preprocessing method combining discrete cosine transform and weaving technology is proposed to improve the key entropy. Third, a highly adaptive quantization method is proposed, including the adaptation of the guard interval, the quantization level and the quantization threshold. Finally, simulation results show that the key agreement and randomness of this scheme is better than general schemes in the same scattering environment, thus verifying the superiority of this scheme.

    Design of the programmable neural network processor based on the transport triggered architecture
    ZHAO Boran;ZHANG Li;SHI Guangming;HUANG Rong;XU Xinran
    Journal of Xidian University. 2018, 45(4):  92-98.  doi:10.3969/j.issn.1001-2400.2018.04.017
    Abstract ( 374 )   PDF (1002KB) ( 74 )   Save
    References | Related Articles | Metrics

    The convolutional neural networks have the problems of structure diversity and large amounts of data exchange and computation. A transport triggered architecture based convolutional neural network processor is presented in this paper. The data transport network is constructed with multi-channel direct memory access channels, the multi-port memory and the specialized pooling data path, which solves the inefficient data exchange problem. Experimental results show that, although the proposed architecture is 11% slower than the streamline structure, it can adapt to a variety of convolutional neural networks and save 46.5% multipliers. Compared with the schemes presented in other papers except pipeline implementation, our design improves the data throughput rate by 40% at least. Besides, this system has advantages of parallel efficiency, programmable flexibility, online architecture reconfiguration, high processing speed, etc.

    Image dehazing by the segmenting median filter and transmission compensation
    HU Yan;WANG Keyan;XU Ning;LI Yunsong;ZHANG Shanshan
    Journal of Xidian University. 2018, 45(4):  99-105.  doi:10.3969/j.issn.1001-2400.2018.04.018
    Abstract ( 363 )   PDF (5837KB) ( 77 )   Save
    References | Related Articles | Metrics

    To enhance the quality of haze-degraded images, a single image dehazing approach based on the segmenting median filter and adaptive transmission compensation is proposed. First, a segmenting median filter is presented to smooth the transmission map estimated by the ‘dark channel prior + guided filter’ dehazing method, which aims to reduce the unnecessary edges or the texture of the transmission map but preserve the edges where the depth changes abruptly. Then, an adaptive compensation scheme is introduced to correct the transmission of bright areas by constructing a compensation function to avoid segmenting sky areas. Besides, we use a simple and effective way to determine parameters automatically, which improves the adaptability of our method. Experimental results show that our method enhances the image contrast and reduces the color distortion of sky areas by exploiting a more accurate estimation on the transmission map. In addition, our method has a better adaptability which can obtain sharper dehazed results both for the images with the sky and for those without the sky.

    Multi-layer incremental feature extraction method for industrial big data
    WANG Xing;HUANG Xiaoyu;LIU Xuanpu;KONG Xianguang;NIU Meng
    Journal of Xidian University. 2018, 45(4):  106-111.  doi:10.3969/j.issn.1001-2400.2018.04.019
    Abstract ( 413 )   PDF (754KB) ( 120 )   Save
    References | Related Articles | Metrics

    We focus on the failure of Incremental Linear Discriminant Analysis in the case of a high-dimensional small sample in industrial big data. An improved multi-layer incremental feature extraction method for industrial big data is proposed to solve this problem which can reduce dimension effectively, and at the same time the variance information and discriminant information on the sample is kept as much as possible. First, the data streams are updated incrementally with the sliding window in real time, and its outliers are detected and filtered. Second, the incremental principal component analysis is made initially to extract the features of the data and the Fisher discriminant function is used to quantify the classification information contained in each principal element. Then the contribution and recognition ability of principal components are weighted by the entropy evaluation method to select the principal components. The selected principal components constitute a new feature space. Finally, Incremental Linear Discriminant Analysis is made to complete the second feature extraction and classification for the high-dimensional data. Experimental results indicate that the improved method could extract the features effectively in real-time and that the industrial data can be discriminated better.

    RNA secondary structure prediction using the quantum genetic algorithm
    LIU Yang;LI Jiaqiao;WANG Fan;WANG Zengbin;SHI Sha
    Journal of Xidian University. 2018, 45(4):  112-117.  doi:10.3969/j.issn.1001-2400.2018.04.020
    Abstract ( 452 )   PDF (660KB) ( 133 )   Save
    References | Related Articles | Metrics

    Due to the high complexity and low search efficiency of the traditional algorithm based on the minimum free energy model, a quantum genetic algorithm for RNA secondary structure prediction is proposed, where the group is initialized by coding with qubits and the corresponding evolutions are accomplished by quantum unitary operations (i.e., quantum gates). By using this strategy, the scale of the required groups is reduced significantly due to the parallelism of quantum computing which also leads to a more powerful searching ability compared with the classical genetic algorithm. Based on the sequences offered by RNA STRAND database, the algorithm was tested by quantum simulations. Numerical results show that, for even only 20% of groups exploited with respect to the classical genetic algorithm, the prediction accuracy yielded by this strategy is still superior to that of the classical one, and that the number of evolution rounds is also obviously reduced by using this algorithm.

    Highly stable FDTD method for complicated dispersive medium
    ZHANG Yuqiang;GE Debiao
    Journal of Xidian University. 2018, 45(4):  118-122+173.  doi:10.3969/j.issn.1001-2400.2018.04.021
    Abstract ( 335 )   PDF (587KB) ( 61 )   Save
    References | Related Articles | Metrics

    A novel finite-difference time-domain (FDTD) approach for analyzing a complicated dispersion model is presented. Starting from the susceptibility in the quadratic rational function form, the explicit FDTD time-step formula is obtained by applying the Newmark algorithm to both sides of the relation equation with the polarization vector and electric field in the time domain. Then, the stability of the presented algorithm is investigated from two aspects of theory and numerical computation. It is observed that this method has the advantages of generality and high stability and can thus be applied to the treatment of many complicated dispersion models, including the complex conjugate pole residue model, the critical point model, the modified Lorentz model, the complex quadratic rational function, etc.

    High-speed high-broadband master-slave sampling and hold circuit
    DING Hao;WANG Jianye;LIU Wei;XIONG Yongzhong
    Journal of Xidian University. 2018, 45(4):  123-128.  doi:10.3969/j.issn.1001-2400.2018.04.022
    Abstract ( 498 )   PDF (1594KB) ( 140 )   Save
    References | Related Articles | Metrics

    This paper presents a new-style high-speed broadband master-slave sampling and hold circuit based on the 0.13μm SiGe BiCMOS. In order to realize DC coupling and sample the low-frequency low-offset-voltage signal, the PMOS source follower is used in the input stage. The Cherry-Hooper structure is used to expand the bandwidth up to 18GHz. Signal feedthrough is cancelled by the master-slave sampling structure and cross-coupled capacitors. Clock feedthrough is attenuated by complementary bipolar transistors. The spurious free dynamic range is 33~38dB. Comparison results show that the proposed circuit has a big advantage in bandwidth and is able to sample at a high sampling rate.

    Multi-authority attribute-based encryption scheme from LWE problem
    YAN Xixi;LIU Yuan;LI Zichen;TANG Yongli;YE Qing
    Journal of Xidian University. 2018, 45(4):  129-136.  doi:10.3969/j.issn.1001-2400.2018.04.023
    Abstract ( 401 )   PDF (542KB) ( 94 )   Save
    References | Related Articles | Metrics

    Most ABE schemes are based on bilinear maps, which leads to a high cost of storage and communication, and which cannot resist quantum attacks. Thus, we propose an LWE-based multi-authority attribute-based encryption scheme. In this scheme, there are different attribute authorities which can manage different attribute sets and generate privacy keys for the users. It can ensure the safety and improve the efficiency of the system. In addition, the Shamir secret-sharing schemes technology is introduced to construct the access tree structure which can support “and”, “or” and “threshold” operations of attributes with a high flexibility. Besides, we use the lattice theory and the basis delegation without a dimension increase algorithm in our scheme, which can not only increase the speed of encryption and decryption, but can also reduce the storage cost of the user′s private key and ciphertext. Lastly, the scheme is proved to be secure against the chosen plaintext attack under the standard model. Compared with the existing related schemes, our scheme can improve prominently the flexibility of the access policy, and it also can achieve a better system performance. Therefore, our scheme is more effective in the large scale distributed environment.

    Effects of complicated interference environment on the outage probability of the frequency hopping system
    LI Jingwei;LIU Guangkai;WANG Chuanchuan
    Journal of Xidian University. 2018, 45(4):  137-142.  doi:10.3969/j.issn.1001-2400.2018.04.024
    Abstract ( 306 )   PDF (585KB) ( 80 )   Save
    References | Related Articles | Metrics

    The frequency hopping communication system is interfered by follower jamming and adjacent frequency interference in tactical applications. Aiming at the influences of both aspects on the frequency system, the mechanism of action is analyzed through theoretical derivation, and the theoretical results are validated by computer simulation. For typical frequency hopping radios, by taking the model above as the parameter of simulation, the outage probability is evaluated under the communication field conditions. The simulation results verify the effectiveness and the correctness of the theoretical analysis.

    Iterative beamforming method in the multi-path channel for the millimeter-wave phased array
    TANG Junlin;ZENG Yuan;YUE Guangrong;LI Shaoqian
    Journal of Xidian University. 2018, 45(4):  143-148.  doi:10.3969/j.issn.1001-2400.2018.04.025
    Abstract ( 371 )   PDF (532KB) ( 80 )   Save
    References | Related Articles | Metrics

    The millimeter-wave array has many antenna elements and the beamforming algorithm in the frequency selective channel generally has a high complexity. In this paper, a low complexity iterative beamforming method in the frequency selective channel for the millimeter wave phased array is proposed. The method first obtains the beamforming vector which is adapted to the millimeter-wave phase array at one end of the communication, and then obtains the beamforming vector at the other end in the same way, and iterates in the communication transceiver side until the algorithm converges. Simulation results show that the beamforming vector obtained by this method can be of high channel capacity and achieves a higher transmission rate, and that it can converge after about three iterations, greatly reducing the overhead of link establishment.

    Multi-OAM-mode travelling wave resonant cavity antenna
    DANG Weiguo;ZHU Yongzhong;SU Daoyi;YU Yang
    Journal of Xidian University. 2018, 45(4):  149-154.  doi:10.3969/j.issn.1001-2400.2018.04.026
    Abstract ( 383 )   PDF (5110KB) ( 132 )   Save
    References | Related Articles | Metrics

    To solve the problems of multimode-working and miniaturization faced with most current orbital angular momentum (OAM) antennas and promote the practical application of OAM in wireless communication, a novel multi-OAM-mode antenna based on the substrate integrated waveguide resonant cavity is designed, fabricated and measured according to the eigen mode theory of the resonant cavity. As simulation and measurement indicate, the proposed antenna successfully resonates and generates OAM with l=1 and 3 on 4.1 and 6.2GHz, respectively, which is in good agreement with theoretical analysis. Compared to existing ones, it is more compact in size and easier to integrate, with a unique ability to generate different OAM modes by multi-mode working of a single cavity, and more practical in wireless communication.

    Multi-mapping convolution neural network for the image super-resolution algorithm
    WANG Shiping;BI Duyan;LIU Kun;HE Linyuan
    Journal of Xidian University. 2018, 45(4):  155-160.  doi:10.3969/j.issn.1001-2400.2018.04.027
    Abstract ( 404 )   PDF (4798KB) ( 209 )   Save
    References | Related Articles | Metrics

    The traditional convolutional neural network for super-resolution will obtain abundant details and edge information with difficulty. By the analysis of the detailed characteristics in three modules of conventional methods, we propose a new multi-mapping convolutional neural network model. By the multi-mapping module, rich and varied characteristics from each layer can be captured. Combining with the error back propagation algorithm, a novel loss function with total variation regularization is used to train and seek optimal parameters, which reconstruct accurate and effective high-resolution images from the network. Extensive quantitative and qualitative evaluations have shown that the proposed algorithm improves effectively the resolution of the image.

    Method for the estimation of the scanning cycle for countering the reconnaissance radar
    MA Xiao;ZHANG Hailin;HAO Benjian;WAN Pengwu
    Journal of Xidian University. 2018, 45(4):  161-165+180.  doi:10.3969/j.issn.1001-2400.2018.04.028
    Abstract ( 589 )   PDF (541KB) ( 125 )   Save
    References | Related Articles | Metrics

    This paper focuses on the problem of accurately estimate the scanning cycle of the reconnaissance radar by using the single one Unmanned Aerial Vehicle (UAV). The Main-Bean Pattern (MBP) of the radar antenna is reconstructed effectively according to the received radar pulses signals based on the norm approximation algorithm. For the traditional Pulse Doppler (PD) radar, the fitting then average method and average then fitting method are proposed to approximate the MBP of the radar antenna. The obtained MBP curve can be utilized to accurately estimate the scanning cycle of the reconnaissance radar. In the simulation section, the validity and the robustness of the proposed methods are verified through the performance comparison. The results indicate that when the receiving signal-noise-ratio (SNR) is high, the average then fitting method can be better close to or even reach Cramer-Rao Lower Bound (CRLB) performance.

    Design of the transcendental function computing unit of the computer vision array processor
    SHAN Rui;LI Tao;JIANG Lin;DENG Junyong;YANG Bowen;SHEN Xubang
    Journal of Xidian University. 2018, 45(4):  166-173.  doi:10.3969/j.issn.1001-2400.2018.04.029
    Abstract ( 377 )   PDF (747KB) ( 76 )   Save
    References | Related Articles | Metrics

    Computer vision algorithms involve a large number of transcendental function calculations. For accelerating the processing speed of the computer vision algorithm on a computer vision array processor and meeting the real time requirement of vision application, an algorithm of piece-wise linear approximation based on adjusting intercept is proposed, and a transcendental function calculator pipeline structure is designed based on the proposed algorithm, which employs a mixture operation system of the fixed-point number and the logarithmic number system. The function of cos/sin is computed approximately. Meanwhile, based on the logarithm/exponent convertor, the other functions are computed in the logarithmic number system. Simulation is done based on hundreds of random data and the architecture is synthesized using SMIC 0.13μm technology. The proposed structure improves the system throughput and reduces the computational error with the clock frequency reaching 200MHz.

    MIMO radar angle estimation for coherent sources under non-stationary noise
    GONG Jian;LOU Shuntian;GUO Yiduo;ZHANG Weitao
    Journal of Xidian University. 2018, 45(4):  174-180.  doi:10.3969/j.issn.1001-2400.2018.04.030
    Abstract ( 320 )   PDF (662KB) ( 114 )   Save
    References | Related Articles | Metrics

    In order to solve the problem of MIMO radar angle estimation under non-stationary noise, multipath effect and coherent electronic interference, an angle estimation method based on oblique projection operator and Teoplitz matrix reconstruction is proposed. Through the reconstruction of Toeplitz, the nonstationary noise is transformed into Gauss white noise, and then the oblique projection operator is used to exclusion of incoherent sources, and the angular estimation of coherent source is realized. This method has no array aperture loss, can be applied to any array structure, and the ability of overload and element saving of the source is stronger. Finally, simulation results indicate that the proposed algorithm is more suitable for low SNR than the multistage Wiener filter algorithm.