2008 Vol. 34, No. 06

Display Method:
Monocular vision-based navigation algorithm for mobile robots in unknown environments
Zhan Qiang, Wu Jia
2008, 34(06): 613-617.
Abstract:
A monocular vision-based navigation algorithm for mobile robots moving in unknown environments was presented, which consists of obstacle detection, monocular vision based distance measurement and local path planning. In order to decrease the impact of lighting disturbance on obstacle detection, pixel-based image segmentation in HSI color space was used to get the set of the image coordinates of the obstacle boundaries by comparing histograms of current image window and reference image window. In order to realize distance measuring between obstacles and a robot, a transformation equation between the vision coordinates and the robot coordinates was deduced by geometrical reasoning. Grid map was used to model the rectangular area transformed from the trapezoidal field of view of the camera and grids with obstacles were marked according to the obstacle boundaries and the measured distance. With a new searching method a safe local path for the robot was found by searching the grid map. Experiments done in real environments show that less sensitive to highlight and shading, the proposed algorithm could correctly plan a safe local path to navigate the robot.
UV treatment of blended epoxy resin and its surface chemical reaction
Yang Guang, Huang Pengcheng
2008, 34(06): 618-621.
Abstract:
The cationic ultraviolet (UV) curing blends of bisphenol A epoxy resin E-44 and silicone-epoxy resin ES-06 were investigated. The UV-curing condition was explored to optimize blended resin formulation. The variations of physical and chemical properties of UV-cured sample surfaces were investigated by scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy (XPS). The results indicate that the photocuring rate of the blended resins is promoted with the increase of E-44 mass fraction in the presence of 5% of a diaryliodonium salt photoinitiator. The gel conversion of a 1∶ 2 mass ratio blend of ES-06 and E-44 reaches 93% after 15-min of UV irradiation. Based on XPS data, it is found that the UV-cured sample surface has been incompletely oxidized by the ozone and atomic oxygen produced by UV radiation from atmospheric oxygen. It may be proposed that the UV-cured materials produce a SiO x film containing carbon atom at the surface of the sample.
Environmental analysis based on PNN and EMD for vibration of spacecraft
Yang Hai, Cheng Wei
2008, 34(06): 622-626.
Abstract:
In view of the disadvantages of the traditional time-varying parameters modeling algorithm about nonstationary random vibration signal of a spacecraft with closed spaced modal frequency, a multicomponent process neural network (PNN) autoregressive model was proposed, which was based on the empirical mode decomposition (EMD). The EMD was utilized to decompose the original time series into several local orthogonal intrinsic mode functions (IMF) with different time scale. The PNN was established for anyone of these IMF, and obtained time-varying power spectrum density (PSD). The time-varying PSD of the original signal was reconstituted by superposing. The simulation and example analyzed results suggest that this method avoids the correlative estimation calculation and reduces the calculation complexity. There is no cross-term interference, so it improves the time-frequency distributing characteristic of the signals. This method has a higher ability of extracting signal characteristic and can be used to analyze the vibration signals of spacecraft under various work conditions.
Model of radar network detection based on average RCS value of different angle territory
Li Ying, Huang Peilin, Wu Zhe
2008, 34(06): 627-629.
Abstract:
There are great differences among the target-s radar cross section values in its different angle territory. A new model about mono radar detect probability and radar network comprehensive detect probability was presented. The model was based on the target-s average radar cross section value in its different angle territory. It adopted two different plans in the computer simulation platform: taking the target-s single average radar cross section value; taking the target-s different average radar cross section in its different angle territory, according to the angle of radar refers to the target. There are differences between the two plans- detect probability results, which prove the model-s rationality. The detect model fits the actual detect situation that radar will detect different radar cross section values from different detect angles. The detect probability result is more accurate which comes from the plan that taking the target-s different average radar cross section in its different angle territory.
Parameter estimation and elimination to fractal noise of fiber optic gyro
Wu Huawu, Ren Zhang
2008, 34(06): 630-633.
Abstract:
There are Gauss white noise and fractal noise in drift of fiber-optic gyro. It is difficult to eliminate this mixed noise using traditional time series analysis method. Based on wavelet analysis, a new parameter estimation approach step by step for fractal stochastic processes and added white noise was proposed. On the basis of the fit for autocorrelation function, the power of white noise was estimated. Then the parameters of fractal noise were estimated in wavelet transform. The suitable threshold was selected to filter fractal noise and white noise of measurements. This estimation and denoising method is effective for this mixed noise in fiber optic gyro which can be proven by experiment results. Further more, a prior knowledge of noise is not needed, which means it is suitable in most situations.
Longitudinal stability of an agriculture aircraft during low-altitude flight operation
Yue Ting, Wang Lixin, Kuang Long
2008, 34(06): 634-637.
Abstract:
The longitudinal aerodynamics of agriculture aircraft during low-altitude flight operation was analyzed. And the longitudinal static stability criterion in ground effect was discussed, including the effects of major aerodynamic derivatives on longitudinal static stability. By the five-order small-disturbance linear equation in ground effect, the influence on the mode properties of agriculture aircraft during low-altitude flight operation was computed. And the approach of mode vectors was used to investigate the dynamic stability, which is useful to understand the major variables acting in different modes. The longitudinal dynamic stability was analyzed by Routh's criteria. To ensure dynamic stability, the favorable position of the center of gravity was deduced by approximations. The results show that the longitudinal aerodynamics and stability influenced by ground effect should not be ignored for the agriculture aircraft during low-altitude flight operation.
Tensile properties simulation of two-dimensional woven reinforced composite laminates after high velocity impact
Zou Jian, Cheng Xiaoquan, Chen Hao, Shao Shigang
2008, 34(06): 638-642.
Abstract:
The tensile performances after high velocity impact of two kinds of two-dimensional woven reinforced composite laminates often used in helicopter structures were analyzed. As for high velocity impact damage characteristics of composite laminates, a finite element model, using commercial software and damage cumulation method, was established to simulate the damage propagation and breakage under tensile load in two-dimensional woven reinforced composites laminates after high velocity impact. The simulation results were compared with experiential formula results to verify this model. It shows that they are well coordinate, and the model is relatively simple and intuitionistic to estimate damage propagation and residual tensile strength, which offers important reference for the two-dimensional woven reinforced composite laminates damage tolerance analysis after high velocity impact.
Rapid frequency acquisition and tracking for pilot pulse carrier
Zhang Jianglin, Lü Shanwei, Dong Shengbo, Wei Zhimian
2008, 34(06): 643-646.
Abstract:
To aim at short-time pilot pulse carrier-frequency synchronization, a rapid acquisition and tracking approach was proposed. In this approach, a phase deduction frequency estimator (PDFE) was employed to perform initial carrier-frequency coarse estimation and accomplish frequency acquisition in short time. The PDFE obtains the phase angle estimation of real signal samples by least square method, and then computes the frequency value by phase unwrapping, unary linear regression and least square fitting. Its estimation accuracy exceeds Prony-s method and its estimation error normalized variance approaches the Cramer-Rao bound of real sinusoid signal at high signal-to-noise ratio. Digital down conversion loop and Kay-s method were adopted to perform frequency tracking, its performance exceeds cross product automatic frequency control (CPAFC) loop. The tracking accuracy was evaluated by simulation. An example was given, its maximum frequency tracking error is not more than 3 Hz at input SNR above 13dB for 1ms pilot pulse width.
Three-dimensional self-localization scheme for wireless sensor networks
Liu Yuheng, Pu Juhua, He Yang, Xiong Zhang
2008, 34(06): 647-651.
Abstract:
A novel three-dimensional (3D) self-localization scheme APIT-3D (approximate point-in-te-trahedron) was presented for wireless sensor networks (WSNs). The target sensor performed APIT-3D test to judge whether it resided inside or outside of tetrahedrons, which narrowed down the possible location areas (PLA). Then, APIT-3D averaged the centers of gravity (COG) of the intersection of PLA as the unknown sensor-s estimated location. Simulation results indicate that, as a range-free scheme, APIT-3D can achieve location accuracy within 40% of radio range, while the communication cost remains reasonable comparing with its 2D counterpart. Since APIT-3D depends on no ranging measures or global infrastructure, it presents a low-cost solution for self-localization of densely deployed and large scale WSNs in 3D spaces.
Variable structure control of testing bed drive for aero hydraulic pump
Bai Guochang, Wang Zhanlin, Chen Juan
2008, 34(06): 652-655.
Abstract:
It is necessary to test the hydraulic power on ground, the method of frequency inverter+asynchronous motor+increasing gear box is one of advanced and effective way in the testing of aero hydraulic pump on ground. The mathematical model of drive system with vector control inverter of aero testing bed was established. Aiming at problems that the pump rotate speed did not keep constant if the load variable at the process of experiment on ground, a control algorithms called sliding mode control(SMC) with integral sliding surface was adopted, and Lyapunov function was constructed to prove the stability. To debase the chatting phenomenon, a control algorithm combine SMC and pseudo derivative feedback(PDF) control supervised by fuzzy control was put forward. The simulation and test results show that this control method has high speed precision and can decrease chatting phenomenon under the condition of retaining the robustness of SMC.
Team autonomous entity path planning algorithm TAEPP
Zheng Yanbin, Zhao Qinping
2008, 34(06): 656-660.
Abstract:
Through analyzing of team autonomous entity path planning problem, a path planning algorithm TAEPP (Team Autonomous Entity Path Planning) was proposed. It is a centralized global planning method. Using mended A* algorithm ,can get path for team leader, and through optimizing,the table of key points and the table of formation changing were achieved, so paths for all non-leader members could be found. Formation changing table made the formation change more reasonable, and decreased the required communication numbers in formation maintaining. Through adjusting the two parameters in the algorithm, different paths emphasized on formation and emphasized on distance would be found. The results of experiment show that TAEPP algorithm can provide more reasonable and more efficacious paths for every member in team, according to the environment and team formations.
Modeling strategy of principle component regression
Wang Huiwen, Wang Jie, Huang Haijun
2008, 34(06): 661-664.
Abstract:
When the mechanism and the reason of failure of the classical principal components regression were analyzed, a new strategy of PCR modeling was presented as:①deriving all components and modeling with all these components; ②exclude all components which were not significant in t-test; ③modeling with the components which were significant in t-test. Proved the regression coefficient and the t-test value of any principal component were unrelated to the other principal components. It was insured that, when applying backward-delete variables law, all the variables which were not significant in t-test test could be deleted together at the same time. It was not necessary to delete them gradually. A simulation study was given to prove the validity of the strategy. The research indicates that the suggested strategy can effectively derive components which are explainable to dependent variables. Modeling under the condition of multicollinearity is enabled, and all the independent variables can be included. The process of suggested variables selection method is simple, and the accumulated error is smaller than that of partial least-squares regression.
Autonomous orbit determination and precise measurement of relative position of formation constellation
Ma Jun, Xiao Yelun
2008, 34(06): 665-668.
Abstract:
A fully autonomous, highly precise method of orbit determination and measurement of relative position for formation constellation was proposed, and the basic scheme, its feasibility and engineering specifications of space absolute orientation——one of the key techniques of autonomous orbit determination were studied. Lastly, validation by using of simulation was carried out. The simulation results indicate that fully autonomous navigation of formation constellation without supports from ground stations and satellite navigation system can be realized by means of highly precise measurement of distance and absolute orientation observation between satellites in the formation constellation; And at the certain error conditions of distance and orientation measurement, the precision of absolute orbit determination can be better than 20m and precision of relative position measurement can be better than 10cm; Relative error has obvious relationship with sampling interval time, and this interval time is recommended to be selected within 10s; Absolute positioning error does not have obvious relationship with sampling interval time, and the main error source is the whole parallel transformation of constellation. The simulation results validate the correctness of proposed scheme.
Novel local route discovery algorithm in ad hoc networks
He Yang, Zhang Ting, Chen Zhenyong, Xiong Zhang
2008, 34(06): 669-672.
Abstract:
In ad hoc networks, to solve the high level of routing overhead problem while on-demand routing protocols using full-net query flooding during route discovery, a local route discovery algorithm was proposed to reduce the routing overhead. Nodes around every route calculated the minimal hop counts to the destination node based on data packets through the route. When one route was broken, the route discovery request packets would be relayed only by nodes around the broken route in the way of the minimal hop counts decreased, which reasonably restricted route discovery region. The route discovery range would be enlarged by adding hop counts of route discovery request packets relayed by nodes around the broken link, which increased local route discovery success rate. The experimental results by simulation indicate this algorithm could contribute to reducing the routing overhead by about 60%, cutting down the average end-to-end delay by about 70%, and increasing the packet delivery ratio by about 10% at most. In addition, any routing protocol depending on flooding based routing discovery can use this algorithm for performance optimization.
New estimation method of target’s radial velocity for stepped-frequency radar
Tian Jinjun, Liu Lin, Xue Minghua
2008, 34(06): 673-676.
Abstract:
A new radial velocity estimation method was presented to compensate the effects of the target-s radial velocity on range profiles synthesized from a stepped-frequency radar. Theoretical analysis indicates that the correlation function derived from two sequential bursts- normalized complex samples of a stepped-frequency radar is a complex sinusoidal sequence, and the sequence-s frequency is related to the target-s radial velocity. So the target-s radial velocity can be estimated with the sequence-s frequency calculated by fast Fourier transform(FFT) algorithm. The estimable velocity range of this method is applicable depends only on the system bandwidth and the pulse repetition period. As the theoretical estimation resolution is determined by the point number of FFT, computational cost can be saved in the case of high estimation resolution is required.Simulation results show that this method can estimate the target’s radial velocity accurately with good anti-noise performance.
Neural network control in active vortex control based on low aspect ratio flying wing
Kong Yinan, Wang lixin, Qian Weiqi, Wang Qing
2008, 34(06): 677-680.
Abstract:
A low aspect ration flying wing aircraft was researched. Full state feedback controller was used for the continuous control effectors to improve the stability of aircraft. To improve the control ability, a vortex flow controller(bang-bang controller) based on three layers BP neural network was used. Through adjust weights for states and the number of hidden layers, the designer could optimize the network. The convergence speed and precision of the network were improved by using variety step network. The evaluation criteria consist of closed-loop system performance and activity level of the vortex flow controller nozzles were given. The results show that application of a neural network controller improves general performance of regular design and makes controlling of the vortex flow controller activity more practical comparing to conventional controller.
Applicability compensation method for random drift of fiber optic gyroscopes
Wang Xinlong, Ma Shan
2008, 34(06): 681-685.
Abstract:
Based on the limitation of founding a time-sequence model and using Kalman filter to wipe off the random noise, an adaptive Kalman filtering arithmetic for on-line compensation of the random errors of fiber optic gyroscope (FOG) was presented. As emerging error in a time-sequence model, the time-varied model and unknown noise can be compensated by introducing a pseudo noise.Consequently, an appropriate high-precision filter was designed for the FOG inertial navigation system. The Allan variance was utilized to analyze the output of FOG, and some main noise of FOG can be extracted and confirmed, then the performance and precision of the time-sequence model of FOG can be evaluated. The practical data of FOG was analyzed. It is shown that the main random noise of FOG are the angular rate ramp, the rate random walk and the bias instability, and the proposed adapted Kalman filter arithmetic can adapt to the time-varied characteristic of the FOG drift, and it is an effective method for FOG to wipe off it-s random drift noise.
Stiffness control technology and application in magnetic rods
Wang Bin, Dong Yunfeng
2008, 34(06): 686-689.
Abstract:
Control torque along geomagnetic vector could not be generated by magnetic rods when they are employed in momentum bias satellites with a pitch dipole moment. Stiffness control policy was developed to provide control torque by using magnetic rods to accelerate the motion of angular momentum vector along orbital normal. It made precession control vector move away from geomagnetic vector and move into location in favor of precession control. The expression of pitch dipole moment for stiffness control policy was developed by adding a stiffness term to classical control policy for momentum bias satellites on sun-synchronous orbit. Precession, nutation, stiffness feedback coefficient and their corresponding ratios were optimized and obtained via ant colony algorithm (ACA) and applied to numerical simulation. The comparison with the performance of the controllers with and without stiffness feedback coefficient indicates that stiffness control policy can improve precession control efficiency and increase pointing accuracy under persistent external disturbances.
Temporal and spatial analysis of infrared images from water jet in frequency domain based on DFT
Gong Weili, Zhao Haiyan, An Liqian, Mao Lingtao
2008, 34(06): 690-694.
Abstract:
Based on 2-D discrete Fourier transforms (DFT) of infrared images from water jet and the analyses of its 2-D spatial spectrum, investigations were made on the spatial scales of passive scalars turbulent flow in high pressure water jet fields, which are represented by infrared images. The characteristic length scales and the law of their temporal evolution process were obtained in the energy (or convection) range, the dissipation range and the inertial range in between. Based on 1-D DFT, the turbulent temporal fluctuation characteristics were obtained for the interest points at large scale eddy center in convection range, sub-scale eddy center in inertial range, dissipation range, as well as in the axis of the jetting flow. In addition, the fractal dimensions were also calculated upon the 1-D spectrum, which was then used for fractal characterization of temporal fluctuations of the interest points in different scale regions of the scalars turbulent flow.
Numerical simulation of ethanol solution heat transfer in rectangular microchannels
Xi Youmin, Yu Jianzu, Gao Hongxia, Cao Xuewei
2008, 34(06): 695-698.
Abstract:
Owing to the factors influencing microchannel heat transfer behavior are complicated, the experimental results were contrary each other. The combination of experiment and numerical simulation has become a valid method to get deep understood of the heat transfer characteristics. On the basis of experimental research, heat transfer characteristic of 30% (volume concentration) ethanol solution in rectangle microchannels was numerically simulated. Two different symmetry structures and thermal boundary conditions of microchannels were used. The fully developed boundary conditions were adopted for the inlet fluid boundary. The results of two different symmetry structures with constant fluid thermal property and that changed with temperature were compared. The results were compared with experimental results as well. It shows that simulation and experimental results are fit well to the microchannels and the work fluid used here with fully developed flow inlet condition and thermal property changed with temperature.
Extraction method of rectangular building based on fussy decision-making
Mao Xia, Zhao Xingyuan, Shen Wei
2008, 34(06): 699-702.
Abstract:
The roof of the building in aerial image is generally rectangle or parallelogram. It makes a significance sense to extract the rectangle accurately and fast, either in military aspect or in civilian aspect.A method based on fussy decision-making for extraction of rectangular buildings from the aerial images was proposed. Firstly, the presumed rectangular space was generated via line feature, which was resulted from beeline-extraction of the aerial image. Then, the support evidences which were get from the image were combined by fussy mathematic method in order to obtain decision-making information.Finally, the generated rectangle was selected and validated according to the decision-making information. This arithmetic uses the non-fixed threshold value to select the supporting evidences to improve the validity. The results of performance assessment also possess the robustness and high extraction rate of the approach developed.
Robust image communication system based on wavelet transform analog coding
Wang Zhe, Zhang Xiaolin, Chen Lei
2008, 34(06): 703-706.
Abstract:
For the applications with complexity and delay limit or in varying channel condition, joint source-channel coding (JSCC) helps to improve the conventional tandem communication system since the integrative system design. For bandwidth and power-limited analog channel, an analog coding and transferring method of image which combines wavelet transform and Shannon mapping JSCC was proposed, and based on more rational distortion estimation in consideration of channel noise, an algorithm for jointly optimum channel bandwidth allocation was developed. Simulations and analysis show that, because of the joint source-channel optimization and the capacity of taking the advantage of analog coding and transferring, the image communication system performs well and has small coding/decoding complexity and delay, and it also exhibits the properties of gracefully degradation and improvement with channel mismatch that are helpful to improve the system robustness and bandwidth efficiency in varying channel condition.
Design of cell image acquisition system based on DSP and FPGA technology
Fan Shangchun, Zhou You, Liu Changting, Wang Junfeng
2008, 34(06): 707-710.
Abstract:
It is a problem to trace and record the growth of the large number of cells manually for a long time. Aiming at the mass data, high sampling frequency, complex caculation and other problems in cell image transmission, a novel method of video image acquisition is described. The key problems in the design of DSP (digital signal processors) system and FPGA (field programmable gate arrays) logic control system and the implementation of JPEG image compression algorithm are also discussed. The system basically consisted of enhanced video input processor, FPGA, DSP and other components. It is favorable for its integrated functions, simple structures and flexible complier. It provides a solution for the image acquisition of long time tracing and recording lagre number of cells, and also has the function of after image processing.
Approach based on compiling optimization and disassembling to detect program similarity
Zhao Changhai, Yan Haihua, Jin Maozhong
2008, 34(06): 711-715.
Abstract:
An approach based on compiling optimization and disassembling was proposed to detect similarity in computer programs. It can detect 12 modification strategies that are often used by students, such as renaming identifiers, adding redundant statements and replacing control structures with equivalent structures. The implemented software, called BuaaSim, translates source code into assembly instructions with the help of compiler and disassembler, removes and replaces those easily changed elements in the assembly instructions, and applies a decision function to calculate the similarity, which doesn-t depend on the order of assembly instructions. A simple clustering algorithm was also introduced to find all groups of similar programs. By using two sets of plagiarized transcripts as testing programs, the comparative evaluation shows that BuaaSim has more advantages than JPlag, a famous similarity detection system.
Fuzzy control design of uncertain nonlinear active fault tolerant control systems
Bai Mingzhen, Wu Huaining
2008, 34(06): 716-720.
Abstract:
By considering the random nature of the failure events and the FDI(failure detection and isolation)decisions in real FTC(fault tolerant control)systems, two Markov processes were introduced to describe the random failures and the FDI decision behavior separately. The Takagi and Sugeno fuzzy model was employed to represent a nonlinear active FTC system with norm-bounded uncertainties. Based on the parallel distributed compensation scheme, a method for the design of a fuzzy control law such that the closed-loop fuzzy system is robustly stochastically stable, was proposed. The sufficient condition for the existence of a robust and fault tolerant control law was given in terms of linear matrix inequalities, which can be solved efficiently by using the existing LMI(linear matrix inequality)optimization techniques. The latter truck-trailer simulation results show the effectiveness of the proposed method.
Lightweight single sign-on service mechanism for portal
Luo Xin, Wu Jing, Xiong Zhang, Pu Juhua
2008, 34(06): 721-724.
Abstract:
An integration framework with easy and efficient implementation is necessary in construction process of information portal. To overcome the limitation that traditional single sign-on mechanisms can-t achieve fast integration in dynamic and loose coupling environment, a novel lightweight single sign-on service mechanism for portal (LSSO-service) was proposed based on service oriented architecture. LSSO-service-s business logic functions were divided according to a distribute integration framework with services higher than object layer. By this means, the simple, adaptive, loose-coupled and flexible single sign-on service could work well, and those application systems based on different techniques might be quickly integrated into portal with a unique entry. The design idea and functional principle of LSSO-service were both presented. To validate the high significance and feasibility of the proposed approach, an integrated prototype was carried out.
Driving torque special solution constructive approach of redundant actuated manipulator
Yan Caixia, He Guangping, Lu Zhen
2008, 34(06): 725-728.
Abstract:
Taking redundant actuated manipulator with planar 3 degree of freedom as subject investigated, the kinematics model was deduced. Based on the virtual work principle and the equivalent force theory, the dynamic model was built. After the fault tolerant performance of redundant actuated manipulator was discussed, the force abrupt change of driving joint before and after fault occurrence was defined. Furthermore, the uncertainty of redundant actuated manipulator dynamics inverse solution was analyzed. To improve the reliability and fault tolerant performance, a new force allocation approach was brought forward. By taking the reliability of actuator as coefficient multiplier, the uncertain solution was constructed from the certain force allocation solution after failure. The sums of force abrupt change got calculated with traditional approach and this new method were compared. Data simulating result indicates this approach has an outstanding fault tolerant performance.
Gram-Schmidt regression and application in cutting tool abrasion prediction
Wang Huiwen, Chen Meiling, Gilbert Saporta
2008, 34(06): 729-733.
Abstract:
Multiple linear regression is one of the most widely applied statistical methods in scientific research fields. However, the ordinary least squares method will be invalid when the independent variables set exists server multicolinearity problem. A new multiple linear regression method, named Gram-Schmidt regression, was proposed by the use of Gram-Schmidt orthogonal transformation in the modeling process. Not only can it screen the variables in multiple linear regression, but also provide a valid modeling approach under the condition of server multicolinearity. The method was applied to the prediction of the flank wear of cutting tool in the turning operation. The results demonstrate that the variable screening is reasonable and the model is highly fitted.
Topic model based structural Web services discovery
Chen Jiangfeng, Yu Jianjun
2008, 34(06): 734-738.
Abstract:
A structural Web services discovery approach based on a topic model was proposed. Each Web service would be modeled as a structural textual document using probabilistic model generated by latent dirichlet allocation (LDA). Every document could be seen as a multinomial random mixture of topics, and each topic had multinomial distribution over keywords, and thus a Web service retrieval based on the topic model was put forward. At the same time, the similar correlations in functionally structural Web services could be modeled as a directed acyclic graph (DAG) and be measured by using a gap-weighted n-spectrum kernel. Finally the experiment results show that several metrics for the classification and selection of services, such as success rate and executes efficiency, were improved via using our approach.