15 January 2018, Volume 44 Issue 1
-
Select all|
-
Computer Engineering. 2018, 44(1): 1-8. https://doi.org/10.3969/j.issn.1000-3428.2018.01.001
A novel binocular Visual Odometry(VO) algorithm is proposed for real-time precise localization of mobile robots.Firstly,it uses accelerated Scale Invariant Feature Transform(SIFT) operator to extract the image features on the left and right image.The sparse stereo matching is carried out after the extracting.In addition,the method of features tracking is applied between the previous and current image.Thus,the initial pose including rotation and translation matrix can be obtained with the motion estimation method based on the RANSAC strategy.Secondly,the image sequence is divided into key frames and non-key frames.In order to decrease the error of the inter-frame motion estimation,a variable sliding window is applied to optimizing the pose of adjacent key frames locally and nonlinearly.Finally,the closed-loop detection is applied by the method of bag of words.Furthermore,all the poses of key frames in the closed-loop are optimized globally to avoid the error accumulation and the drift of the trajectory.Experimental results show that the proposed algorithm has good real-time performance,while reducing the position pose error and improving the positioning accuracy.
-
Computer Engineering. 2018, 44(1): 9-16.To solve the navigation problem of mobile robot in unknown environment,based on cognitive development mechanism of biology,a cognitive development model is constructed for mobile robot autonomous navigation.The neural network which can dynamically adjust the structure is designed by autonomous inserting neuron.It is used to imitate the properties of biological development and to obtain the network that can match the application requirements.The asymptotic learning characteristic is imitated through the thermodynamic process and a cognitive learning algorithm is designed,which is proved to be convergence in theory.The results of experiments show that the proposed model can make the robot obtain knowledge automatically and accumulate experience from environment like animal,and learn the skill of autonomous navigation through cognitive development.
-
Computer Engineering. 2018, 44(1): 17-22.Aiming at the problem that the geometric parameters are not accurate and the absolute accuracy of the industrial robot is low,a geometric parameter calibration method based on Pose Modify Position Sensitive Detector(PMPSD) is proposed.Through the establishment of the error kinematics model,and the PSD device is used to sample the data,the pose correction principle is used to correct the end laser pose and joint angle,and the model constraint objective function is established.The geometric parameters errors are obtained by using the LM algorithm,and the nominal values of geometric parameter are corrected by the parameter errors.Experimental results show that the method avoids PSD feedback control,and can quickly realize geometric parameters calibration of industrial robots,the positioning mean error and standard deviation are 78.28% and 76.38% respectively,it effectively improves the positioning accuracy of the robot.
-
Computer Engineering. 2018, 44(1): 23-29.In order to make the indoor mobile robot better build the accurate map from the data of 2D laser range finder with noise,an improved algorithm based on similar triangules denoising rule is proposed.Using the splitting algorithm to extract the rough line set from the preprocessing data,the data points between the two split points are denoised by improving the similar triangules denoising rule,and the denoised data are re-split,and scanning points are fitted in the data between each two split points by least square method.Experimental results show that the proposed algorithm can reduce the number of effective points which are eliminated,and the accuracy and false positive indexes are better than similar triangles denoising method and traditional split and mergeing algorithm,at the same time,the line segment merging process is basically avoided,and the robustness and accuracy of environment modeling are enhanced.
-
Computer Engineering. 2018, 44(1): 30-34,43.
The stability control strategy of biped robot walking process is studied,and a crus vibration control system based on the vibration acceleration of the upper body of the robot as feedback is designed.Through virtual simulation analysis software——Automatic Dynamic Analysis of Mechanical System(ADAMS),the virtual prototype of the biped robot is built and imported into Matlab.The automatic anti-disturbance control algorithm is designed to suppress the low frequency and high frequency according to active and passive vibration attenuation.Simulation results show that active and passive vibration attenuation greatly reduces the vibration of the upper body of the robot and compensates the situation that passive vibration can not suppress the low frequency vibration,and effectively improves the stability of the robot.
-
Computer Engineering. 2018, 44(1): 35-43. https://doi.org/10.3969/j.issn.1000-3428.2018.01.006In large-scale temporal documents similarities and differences do not have the ability to identily topics from temporal documents and to track and analyze topics over time.To this end,a method of topic change detection for temporal document corpus is proposed.Similar topics and similarities and foundations are found in the temporal document corpus.Using the improved joint Nonnegative Matrix Factorization(NMF) algorithm,similarities and differences were found in the the timeseries document.To avoid the introduction of noise topics,by calculating the topic of all topic entropy,in order to obtain high-quality topics.Use the word cloud and trend graph to analyze the trend of topic change.Experimental results of two real data sets,20Newsgroups and LTN2011 show that this method can effectively find similarities and differences from the tempord of documents,and the extraction topic is effect and the accuracy is high.
-
Computer Engineering. 2018, 44(1): 44-50.
The function identification of Public Bicycle System(PBS) plays an important role for the balancing scheduling and layout planning of the system.As a new public transport mode,PBS has accumulated more and more data,which not only can reflect the social and economic activities of users at different times and locations,but also has a close connection with the function of the rental points.By combining the spate-time attributes of the leased point and through the historical data analysis of the system,this paper constructs a rental points clustering model of public bicycle system.It makes use of the Latent Dirichlet Allocation(LDA) model and K-means clustering algorithm to find the functional areas of the system,and through the analysis of the clusters’ use features and uses Point of Interest(POI) data and station names to verify the results.The results show that the proposed model can help the system managers grasp the function distribution of the rental points in PBS.
-
Computer Engineering. 2018, 44(1): 51-55.The traditional outlier detection algorithm based on Euclidean distance can not guarantee the accuracy and the running time is too long in high-dimensional data detection.Based on the characteristics of high-dimensional data flow,an improved outlier detection algorithm based on angular variance is proposed by using the method of angle variance.The optimal data set grid and the nearest data grid are constructed to calculate the small scale data flow to measure the abnormal degree of the latest data points,and the improved algorithm is used to the elevator real data flow detection in wireless sensor network acquisition to achieve elevator fault detection.Experimental results show that compared with ABOD and HODA algorithms,the improved algorithm can effectively identify abnormal points in high-dimensional data streams and can be applied to high-dimensional data streams with high real-time requirements.
-
Computer Engineering. 2018, 44(1): 56-61,68.The time series data contains trend information,which can extract the trend turning point according to the trend information of the data,and can achieve the purpose of compressing the data and reducing the influence of noise.By analyzing the trend information of time series data,an adaptive data trend turning point extraction algorithm is proposed.The algorithm does not rely on any prior knowledge,only according to the trend characteristics of the data itself automatically extract the trend turning point,extracted information including the coordinate index and the corresponding data.Compared with SEEP,CAP and PAA algorithm,experimental results show that the fitting error and classification error rate of the algorithm are smaller in the case of multiple data,and the average fitting error is 0.373 6,the classification error rate compared with the original data classification error rate decreases by 3.39%.
-
Computer Engineering. 2018, 44(1): 62-68.Collaborative filtering and probability model are commonly methods for location recommendation.However,the former does not consider mobility patterns of users and the latter performs poorly in sparse dataset.For lack of existing methods,this paper constructs a framework based on Gaussian Mixture Model(GMM) in sparse data,named GMMSD.The check-in history data is divided by time slot,then the user-region matrix is obtained by data preprocessing and the accuracy of recommendation is improved by Matrix Factorization (MF)algorithm in sparse data.Finally,the GMM is learned to predict the probability distribution of different locations where each user checks in.Experiments are carried out on real data sets.The results show that GMMSD can effectively improve the accuracy of location recommendation in sparse data compared with the traditional methods.
-
Computer Engineering. 2018, 44(1): 69-73,78.Supervised-Latent Dirichlet Distribution Allocation (s-LDA) model cannot handle the multi-label problem and s-LDA model is not correct distribution in the classification model.The Supervised Labled-LDA(sl-LDA) model is proposed by adding a category label based on the response variable.It analyses s-LDA model and existed problem of topic classification,through verifying the classification accuracy of sl-LDA model,the paper classifies the sl-LDA model and s-LDA model.Experimental results in the Chinese and English news corpus show that English corpus classification performance is improved by about 3.80% and Chinese corpus is improved by about 1.77%.
-
Computer Engineering. 2018, 44(1): 74-78.In order to solve some problem of dynamic vehicle routing,this paper proposes a vehicle routing algorithm based on big-data calculation model.This algorithm uses big-data calculation to collect and process the initial data,it sets up the alerts trigger in key routing nodes to implement routing mechanism update dynamically,and transforms traditional single time cost vehicle routing problem into transportation cost optimization problem based on dynamic time sequence,and gives the total transportation cost of objective function.Numerical simulation results with the Sioux Falls network show this algorithm is quite effective for vehicle routing problem,compared with the traditional PFIH algorithm and Tabu algorithm,the time performance has obvious advantages.
-
Computer Engineering. 2018, 44(1): 79-83,90.
Before using D-S(Dempster-Shafer) theory for evidence fusion,the extent of the conflict among evidences should be determined.Aiming at this problem,this paper proposes a conflict measure and correction method based on trust degree and false degree.It calculates both trust degree and false degree of each evidence and sorts them respectively.The evidence with smaller trust degree and larger false degree is called high conflict evidence and done weighted correction.Then the Dempster combination rule is used for evidence fusion.Experiment results show that the proposed method can correctly judge the high conflict evidence and solve the problem that the fusion result contradicts the fact,while improving the speed and precision of convergence.
-
Computer Engineering. 2018, 44(1): 84-90.Aiming at the problem that the subjective and objective weights can not be effectively fused in combination weighting,a method of determining the weight of indexes based on the priority degree of adjacent target and the maximum entropy criterion is proposed.The subjective weights of adjacent objects are used to determine the order of importance of the index,the objective weight of the subjective weights is obtained by using the maximum entropy criterion weighting method,and the objective weight,which is closest to the subjective weight of the index,is found by using the vector similarity principle as the final weights.Combining with the case of determining the index weight of commander’s decision-making effectiveness evaluation,the method is proved to be reasonable and effective through comparative analysis of different methods.
-
Computer Engineering. 2018, 44(1): 91-97.The traditional enterprise heterogeneous system integration approaches have the disadvantages such as poor scalability,poor interoperability and high integration cost.It is difficult to meet the requirements of the aircraft support information system integration.In this paper,the heterogeneous system integration problem is researched based on the Service-Oriented Architecture(SOA) and Web Service,and an aircraft support information system integration model is constructed,which includes the service provider module,service registration agent module and service requestor module.In the service provider module,the Web Service Description Language(WSDL) file is used to describe the service,then the service logic processing,data encapsulation and transmission are realized by using the Spring-WS framework.In the service registration agent module,the service registration and response to the query service are completed.The service requester module involves the binding and invocation of the service.Application result shows that the integration system can realize the sharing of bussiness and data in heterogeneous system,which verifies its feasibility.
-
Computer Engineering. 2018, 44(1): 98-103.The core of the shunting engine path planning is to arrange the moving path of each group (disintegration) to complete the grouping(disintegration) task.But the current preparation work is still man-mode,there is big workload,planning is not timely and difficult to adjust.Therefore,a logic programming method based on Answer Set Program(ASP) and Constraint Logic Program over Finite Domains(CLP(FD)) is given.In the case of increasing orbital scale,the time efficiency of the two methods in solving different problem scales and the operation of the corresponding solver are discussed.Experimental results show that ASP is better than CLP(FD) in solving this kind of problem,and the results of path scheduling can be obtained quickly.The corresponding DLV solver can overcome the influence of increasing or decreasing the number of tracks.
-
Computer Engineering. 2018, 44(1): 104-109,115.Traditional semi-physical simulation test platform has problems such as low distributed performance and inconvenient interface expansion.In view of this,this paper adopts ‘layering+distributed’ design concept and designs a semi-physical simulation test platform.It adopts the distributed architecture and decomposes tasks such as test task scheduling and incentive data generation into the test controller,then makes the platform support the simulation and extension of various bus interfaces by means of a domain protocol library.Application and real-time evaluation results show that the proposed platform has good universality and high test efficiency.Its real-time performance can reach hundred micro seconds.
-
Computer Engineering. 2018, 44(1): 110-115.Due to the diversity of mobile devices and the heterogeneity of input and output channels,it is difficult for multiple devices to develop a unified interface description.As a result,the same application usually needs to develop multiple user interfaces for different device characteristics.Aiming at this problem,this paper proposes a design method for mobile user devices based on abstract device.It abstracts the characteristics of current mobile devices and proposes the concept of the abstract device.Then the abstract user interface model can be described in the multi-device environment by the definition standard of the abstract device.Experimental results show that the proposed method can make interface designer only need to design a unified mobile user interface description on the abstract device and finish the generation of the final interface on many different types of mobile devices through mapping.It also can reduce the learning cost and shorten the development cycle.
-
Computer Engineering. 2018, 44(1): 116-120.In order to improve the efficiency of the wideband jamming signal sampling and processing of hardware,a novel method is proposed for parameter estimation of wideband Linear Frequency Modulation(LFM) jamming signal.When priori knowledge is unknown,a narrow-band filtering is used for the wideband LFM jamming signal,whose frequency-modulated rate is estimated by the rapid linear frequency modulation technique.Based on that estimated rate,the wideband LFM jamming signal is demodulated by down frequency demodulator.The frequency shifty of jamming signal will be eliminated and the demodulated signal becomes spare in the discrete Fourier domain.The demodulated signal is compressed sampled and partially refactored for parameter estimation.Simulation results show that the proposed method can accurately estimate the parameter of the wideband LFM jamming signal,even with the points of random projection far less than that of Nyquist sampling.
-
Computer Engineering. 2018, 44(1): 121-127.In order to solve the design problem of predictive Transmit Antenna Selection(TASP)/receive antenna Maximum Ratio Combining (MRC) wireless communication system with the Minimum Mean Square Error(MMSE) Wiener channel prediction,the exact and approximate expressions for the Average Symbol Error Rate(ASER) of the Rectangular Quadrature Amplitude Modulation(RQAM) and Differentially Encoded Quadri-Phase Shift Keying(DE-QPSK) are derived by using moment generating function method and approximate expression of Gaussian Q-function,respectively over time-selective Rayleigh block fading channel.The numerical and simulation results verify that the exact and the approximate ASER analysis of RQAM and DE-QPSK is correct and accurate.The obtained exact and approximate expressions for the ASER of RQAM and DE-QPSK modulation can be used to evaluate the impact of transceiver antenna number,channel predictor parameters and various channel parameters on the ASER performance.
-
Computer Engineering. 2018, 44(1): 128-133.The sub-matrices on both diagonals of the right half of the check matrix are unit matrices.Existence of the deterministic unit matrix not only destroys the randomness of Quasie Cyclic-Low Density Parity Check(QC-LDPC) code,it leads to a certain loss of performance.For this reason,a construction method of QC-LDPC codes with girth 8 and a fast encoding is proposed.An improved quasi-doubly-diagonal parity check matrix is obtained by using Independent Row-column Mapping Sequence(IRCMS) algorithm,row and column rotation shift and masking techniques,so that the constructed codes not only have a girth of at least 8,but also use check matrix to fast coding,they can effectively reduce the coding complexity.Simulation results show that compared with the rule-based codes constructed based on IRCMS algorithm,the constructed codes not only have fast coding characteristics,but also have a performance improvement of about 0.15 dB.Compared with the QC-LDPC codes constructed based on the Progress Edge Growth(PEG) algorithm,the constructed codes are similar in performance with low coding complexity.Compared with the improved DVB-S2 codes,the constructed codes have a coding gain of about 0.1 dB.
-
Computer Engineering. 2018, 44(1): 134-138,143.In the Cognitive Cooperative Radio Network(CCRN),when the secondary users blindly to participate in relay primary users,they will choose the primary users that is not conducive to energy efficiency of the system,thus reduces the overall energy efficiency of secondary user system.To solve the problem,this paper proposes a resource joint allocation scheme.The secondary users have the right to reject the assistance applications that are not conducive to the performance of system,and select primary users who are beneficial to their energy efficiency to assist.In the case of determining the primary user set,through the effective allocation of bandwidth and power,the energy efficiency of the secondary user system is increased.Simulation results show that the scheme can improve the efficiency of secondary system in the condition of ensuring the quality of service for all users.
-
Computer Engineering. 2018, 44(1): 139-143.Ant cluster network architecture of multi-space-time,high-dimensional stochastic process is built based on signal detection and tracking metadata fants dustering algorithm,so that any ant individuals according to the random access signal opportunity to join the transmission path in order to reduce external interference and offset signal distortion.Based on the propagating process of GPS signal frequency domain representation of the opportunistic ant is cluster network architec ture,the generation and fusion of multi-frequency GPS weak signal body are given.A new multiband fusion GPS weak signal tracking and acquisition algorithm is proposed.Simulation results show that the proposed algorithm has obvious advantages over real-time ant clustering control,GPS weak signal acquisition sensitivity and tracking accuracy compared with the tracking acquisition algorithm without extracting weak GPS signal.
-
Computer Engineering. 2018, 44(1): 144-148,153.In the process of Wireless Sensor Network(WSN) node mobile localization,the network topology structure is updated constantly,which affects the accuracy of the predicted model.This paper proposes a mobile localization algorithm based on Laplacian mapping,analyzes the layout characteristics of WSN,introduces local topology information and non-beacon node information,and studies the node localization problem in the semi-supervision framework.Simulation results show that compared with similar algorithms,this algorithm improves the generalization ability of predicted model,increases the localization accuracy of nodes,and has higher stability.
-
Computer Engineering. 2018, 44(1): 149-153.This paper investigates the security performance in the full-duplex relay network.Targeting on the improvement of the security performance,the power allocation is optimized with the consideration of the maximum transmit power and minimum security rate constraints.An efficient iterative algorithm based on DC programming and convex optimization is proposed to solve the nonconvex and difficult optimization problem.Simulation results verify the power allocation strategy proposed can achieve a higher security rate than the conventional average power allocation strategy.
-
Computer Engineering. 2018, 44(1): 154-159.In order to solve the problem that the existing group authentication schemes cannot identify the specific illegal participant,this paper proposes a centralized (t,m,n)-AS group authentication scheme based on Shamir secret sharing mechanism.First of all,it verifies whether all m participants are legal within O(1) time through the overall authentication.Then the scheme identifies all possible adversaries within O(m) time through single authentication in case of illegal participants.Analysis results show that the proposed scheme can resist the conspiring attacks of up to (t－1) members and prevent a non-member adversary with at most (m－1) intercepted tokens from passing authentication.In addition,it does not depend on any hard problem,which is theoretically secure.
-
Computer Engineering. 2018, 44(1): 160-164.
Traditional leader based and decentralized algorithm for solving Byzantine consensus exists the problem of low fault tolerance and high message complexity during the process of solving legitimacy verification.In order to solve these problems,this paper proposes a new blockchain consistency algorithm.It introduces two-phrase commit and quorum voting,using the characteristics of distributed ledge in Blockchain protocol to solve legitimacy verification and ultimate consistency is proved later on.Experimental result shows that compared with traditional Byzantine consensus protocol,this algorithm reduces the complexity of message passing and improves the system fault tolerance rate.
-
Computer Engineering. 2018, 44(1): 165-170.Because some network anomalies have little effect on traffic flow,it is difficult to find such anomalies in traffic analysis.Traditional anomaly traffic detection method based on Principal Component Analysis(PCA) is not suitable for continuous local anomalies detection,and it can reduce the detection accuracy of abnormal flow and the physical meaning is not clear.Aiming at the above situation,an anomalous traffic detection method based on Multidimensional Entropy-Projected Gradient Non-negative Matrix Factorization (ME-PGNMF) is proposed.Firstly,the network traffic data is processed into multidimensional entropy matrix,then Projected Gradient Non-negative Matrix Factorization(PGNMF) is used to reconstruct the multi-dimensional entropy matrix,and the normal subspace and abnormal subspace are separated.Finally,the anomaly is detected by multivariate statistical process control chart Q. Experimental results show that the proposed method can detect the continuous anomaly faster and more accurately than the traditional Nonnegative Matrix Factorization(NMF) method based on the PCA method based on the flow analysis.The low-speed Distributed Denial of Serviee(DDOS) attack anomaly detection is not sensitive to the traffic change.Attacks are more sensitive.
-
Computer Engineering. 2018, 44(1): 171-175,181.Existing permission assignment constraints pay more attention to ensure the safety,however the availability for the system is ignored.Aiming at this problem,this paper proposes a High-level Permission Assignment (HPA)constraint to satisfy the safety and availability requirements simultaneously.It defines the safety verification problem and consistency verification problem,which determine whether an access control state can satisfy a HPA constraint,and whether there exists a valid access control state to satisfy a set of high-level permission assignment constraints.The computing complexities of the safety verification problem and the consistency verification problem show that computing complexities of these two problems are NP-complete and NPNP.It also designs an optimized algorithm for the consistency checking problem by incorporating preprocessing and reducing to Satisfiability(SAT) problem solver.Simulation experimental results verify the effectiveness of the proposed algorithm.
-
Computer Engineering. 2018, 44(1): 176-181.Most of the existing anonymous algorithms only implement the privacy protection by quasi-identifier generalization,which do not consider personalized protection issues of the sensitive attribute.Aiming at this problem,by using the p-sensitive k-anonymity model,this paper designs a personalized privacy protection algorithm based on sensitive attribute.It defines sensitive levels of sensitive attributes based on the user’s own sensitivity and uses sensitive attribute generalization tree to publish low accuracy sensitive attribute values,so as to realize the personalized protection of sensitive attribute.Experimental results show that the proposed algorithm can shorten the execution time and reduce the amount of information loss,meanwhile meeting the requirements of sensitive attribute personalized protection.
-
Computer Engineering. 2018, 44(1): 182-186,192.In physical layer secure communication system,the existing retransmission technology retransmits the whole wrong message when the legitimate user is in error,which may lead to the eavesdropper receives a lot of useful information.Aiming at solving the above problems,the Reliability-based Hybrid Automatic Repeat Request (RB-HARQ) protocol is introduced in,and a secure communication technology is proposed.RB-HARQ protocol can just retransmit the unreliable bits required by the legitimate user and minimizing the information leak to eavesdropper,thus secure and reliable communication can be achieved.Simulation results show that,compared with the traditional HARQ protocol,RB-HARQ protocol can enhance the physical layer security performance of the communication system.
-
Computer Engineering. 2018, 44(1): 187-192.At present,Bayesian structure learning algorithms commonly have the disadvantages of poor convergence,low accuracy and easily trapping in the local optimum.A novel structure learning algorithm is proposed in this paper,which retained the sort framework of Moth-flame Optimization(MFO),defines the crossover operator and variation operator by borrowing the ideas from genetic algorithm to replace the location update strategy of MFO.The mutual information between nodes is considered during mutation period to increase the possibility of returning a similar solution.Simulation results show that the novel algorithm can learn the optimal Bayesian network structure quickly and the returned outcome is more approximate to the standard networks.
-
Computer Engineering. 2018, 44(1): 193-198.The specific practices of this model are that:firstly,seven feature variables are selected for travel mode detection based on the travel characteristics of university students;afterwards,six travel modes (walk,bicycle,electric bicycle,campus bus,bus and taxi) university students selected commonly are selected;finally,IPSO-SVM is used to identify six selected travel modes.This model is using IPSO to optimize SVM parameters,and a travel mode identification method of university students is proposed.Experimental result shows that the average detection accuracy of the proposed method is 94.22%,higher than that of BP neural networks,the decision trees,support vector machine and particle swarm optimization-support vector machine.
-
Computer Engineering. 2018, 44(1): 199-205.In order to improve the performance of Primi speech recognition system,the deep learning model is introduced into Primi speech recognition.The deep learning model is a large capacity and complex network model.Kaldi speech recognition toolkit is used as an experimental platform and five different acoustic models are respectively trained which contain a deep neural network model with four hidden layers.By comparing the speech recognition rates obtained by different acoustic models,it is found that the G-DNN model improves the accuracy of speech recognition by 49.8% than the Monophone model.Experimental results show that the Primi speech recognition rate based on the deep learning model can be improved,when the number of Primi speech corpus in the training set is increased.And the robustness of the Primi speech recognition system based on deep learning is stronger than the other four acoustic models.
-
Computer Engineering. 2018, 44(1): 206-210.Against the problems that most of the existing single attribute recognition is limited and the subjectivity during attribute selection of multiple attribute recognition in social network,this paper proposes a perspective method for core node recognition.Based on optimizing multiple objective of interest,it combines the analysis of social network and system science.Then,combining with multiple classical decision attributes,it uses reevaluation to improve the accuracy of the algorithm.Experimental results in the ARPA and Zachary’s karate club network show that the proposed method can improve the recognition rate of core node sets in social network.
-
Computer Engineering. 2018, 44(1): 211-218,225.This paper presents a parallelizing discrete Glowworm Swarm Optimization algorithm based on MapReduce (MRDGSO) to solve the large-scale Web service selection problem in cloud computing environment.The algorithm redefines individual coding,calculates the distance between individuals,and improves their position update,which increases the search capability in the high-dimension space.The optimization is based on the idea of divide conquer and group and ideal point to avoid falling into local optimum,and improve the ability to deal with large scale problems.Experimental results indicate the feasibility and efficiency of the presented algorithm to solve Web service selection problem and demonstrate the superiority of the algorithm in the scalability.
-
Computer Engineering. 2018, 44(1): 219-225.In order to research the web page click stream data of user’s,mining user interest to recommend personalized learning resources for them,this paper proposes the JMATRIX algorithm.Based on the user’s historical resources click stream information,setting up the directed-graph model of user’s resources click data,and transforming the directed-graph model into matrix model to store.By solving the similarity of matrix model,to obtain the similarity of users,it greatly reduced the complexity of solving user’s similarity of resource click frequency and resource click path,and improved the efficiency and accuracy of the user’s similarity.Combining the Leader Clustering algorithm and rough set theory to realize the user personalized resources recommendation.Experimental results show that the JMATRIX algorithm has higher efficiency and more accurate recommendation effect compared to Leader Clustering algorithm.
-
Computer Engineering. 2018, 44(1): 226-232,237.To solve the problems of the traditional collaborative filtering algorithm which in the process of users’ similarity calculation that due to data sparsity the traditional similarity could not be used in calculating the similarity and cause the information distortion and virtual height,this study proposes a new algorithm that combine the discrete contents with the interests appropinquity degree.Collects the users’ ratings at the beginning of this study,that analyzes the ratings from the aspect of all users’ information ratings and calculates the similarity between user rating vectors by introducing discrete contents theory,conducts a preliminary screening of the similar users,the similarity results are further weighted by the users’ interests appropinquity degree,in the final of this study the users’ interests information fuses in the similarity results,on this basis,the personalized recommendation that is proceeding well by using collaborative filtering algorithm.Experimental results show that the algorithm can significantly improve the quality of recommendation,at the same time,it can maintain well performance in the case of extreme data sparsity.
-
Computer Engineering. 2018, 44(1): 233-237.
In order to segment interested objects in image accurately and quickly,an improved Wolf Pack Algorithm(WPA) is proposed combined with 2D maximum entropy to realize the effective segmentation of image targets.In travel link of WPA,the chaotic systems is used dynamically to adjust the inertia weight.In the attack link,chaotic global search is finisned in the whole solution space.The criterion function is maximized by using the improved WPA combining the 2D maximum entropy.Experimental results show that the proposed algorithm can get accurate image segmentation effect.Compared with the 2D maximum entropy combining with the origin WPA,it can improve the speed and accuracy of image segmentation.
-
Computer Engineering. 2018, 44(1): 238-246.Most of the research on visual saliency in computer vision is limited to 2D image,and ignore the fact that human vision attention decision making operates in real 3D dynamic scene.So this paper proposes a 3D vision saliency algorithm which combines saliency computing framework with multiple features.Using color information of the scene,motion information and depth information,the saliency results of each characteristic channel are calculated respectively,and the final saliency results are obtained by dynamic fusion.Due to the lack of datasets of saliency in 3D scenes,this paper also proposes a new dataset for the evaluation of relevant algorithms.Compared with HC algorithm,RC algorithm,and GMR algorithm,quantitative results prove that the proposed algorithm outperforms all the other algorithms and better fits the visual attention mechanism of human eyes.
-
Computer Engineering. 2018, 44(1): 247-251.Body part classification is an important precondition and crucial step in human pose tracking.Although the random forest classical method with gradient feature of depth can get the request of real time,it remains the weakness of low accuracy,not robust enough to the noise and huge memory consumption.This paper proposes a new classification method,which combines the classical depth feature with RGB edge feature,and introduces the random jungle model to solve the increased dimension of the feature.Experimental results show that the proposed method not only saves about 20% running time,but also improves the test accuracy of about 1%.
-
Computer Engineering. 2018, 44(1): 252-257.In terms of the loose strands problem of electric potential line in the catenary support and suspension devices of high-speed railway,an automatic electric potential line loose strands detection method based on machine vision technology is proposed.The Histogram of Oriented Gradient(HOG) characteristics of the part are extracted.The hybrid classifier of AdaBoost cascade and Support Vector Machine(SVM) are trained to realize electric potential line location.A criterion for evaluating electric potential line faults is given by measuring the area of electric potential.Experimental results show that the proposed method has high detection accuracy.The method can reduce the intensity of catenary operation and maintenance work significantly and has a certain popularization value.
-
Computer Engineering. 2018, 44(1): 258-262.
In order to improve the precision rate of video smoke detection for wildland-urban interface,a video smoke detection algorithm based on multiple image features and deep learning is proposed.Firstly,ViBe(Visual Background extrator)method is applied to extract the moving or changing foreground areas in surveillance videos and image corner information is used to exclude the disturb from objects with detail texture.Secondly,color space features are utilized to narrow the check area for smoke.Thirdly,the cumulated difference image is calculated to find and eliminate the influence of rigid bodies.At last,the deep learning model is used to recognize the filtered image area.The algorithm is designed in the framework of cascade classifier and implemented by using parallel computing technology.Experimental results and project cases verify that the proposed algorithm is significant for improving the precision rate of smoke detection for wildland-urban interface.
-
Computer Engineering. 2018, 44(1): 263-267.In order to solve the problem of the irreconcilability between invisibility and robustness of watermarking algorithm,a novel watermarking algorithm using Non-subsample Shearlet Wave Transform(NSST) is proposed.The algorithm uses shearlet wave transform to images.The subband with strongest texture directionality after NSST is selected to carry watermark based on Singular Value Decomposition(SVD).Experimental results show that the Structural Similarity Index Measurement(SSIM) is above 0.99.The watermark can be extracted and the Normalized Correlation(NC) reaches 0.98 on average after a series of attacks,which shows that the performance is greatly improved.
-
Computer Engineering. 2018, 44(1): 268-273.A modified implicit type Landweber image reconstruction algorithm is presented for solving the ‘soft-field’ nature and the ill-posed problem in Electrical Capacitance Tomography(ECT)technology.On analysis of the basic principles of the ECT system,the paper proposes a formula of the modified implicit type Landweber algorithm to solve the problem of electrical capacitance tomography,and finishes the iterative modification,and analyzes the convergence of the algorithm.Simulation experimental results show that compared with Landweber,LBP methods,both precision and speed of the modified implicit formula Landweber method are very good,and it is simple and stabilized.
-
Computer Engineering. 2018, 44(1): 274-279.
The state detection method of eye and mouth is the key issue for fatigue detection,but it is affected by changing of illumination and wearing glasses.To solve above problems,a fatigue detection method based on facial behavior analysis is proposed.It designs an infrared video acquisition system for driver.The driver’s face is detected based on AdaBoost and the Kernelized Correlation Filter(KCF) tracking algorithm.The feature points are determined by the method of cascade regression,and the eye and mouth regions are obtained.Convolution Neural Network(CNN) is utilized to detect the state of eye and mouth.On this basis,the fatigue parameters are calculated for fatigue detection.Experimental results show that the method can detect the state of eye and mouth accurately and detect fatigue more effectively in many circumstances.
-
Computer Engineering. 2018, 44(1): 280-284.In the Philips audio fingerprint retrieval algorithm,the construction of a query table as an index,which limits the application due to excessive memory consumption.For this reason,an improved algorithm is proposed based on Philips audio fingerprint retrieval.Combining the Fibonacci number and the right shift operation,a new hash function is constructed.The distribution of hash value is optimized by Fibonacci.The length of the hash table is adjusted by performing the right shift operation.Experimental results show that the improved algorithm can reduce the memory consumption and improve the practicability of the system.
-
Computer Engineering. 2018, 44(1): 285-291.For Wireless Sensor Network(WSN) protocol with complicated communication mechanism and abundant supported function as ISA100.11a,traditional design of endpoint based on embedded single chip computer system is hardly to meet practical requirements.Hence,a design of endpoint in which the User Application Process(UAP) is separated from other protocols is proposed.Combining the object oriented structure of the application layer of ISA100.11a protocol,the user application processes and specific I/O interfaces of the application layer are defined as the application module,and the other parts of the protocol are defined as the communication module.One side,limited resource of single chip is no longer needed to deal with application and communication functions simultaneously to avoid bottleneck of performance.On the other side,flexible and extensible I/O interfaces which can be collected and managed are provided,then on this basis the tunnel transmission function in the ISA100.11a protocol which can be compatible with data from outer protocols can be implemented more conveniently.Experimental results show that the design can reduce the rate of packet missing of the whole WSN system efficiently,and provide compatible capacity of data from outer protocols such as ZigBee.
-
Computer Engineering. 2018, 44(1): 292-298,305.To solve the problem of cascading failures in smart grid,this paper proposes a continuous cascading failure simulation model base on the network topology and physical characteristics.The smart grid is abstracted into weighted network according to the active power flow value of the grid line,and carry out dynamic simulations of cascading failures under different removal strategies with using network efficiency to measure the overall performance of the network.The research shows that the larger the edge weight,the greater the impact of the removal of the edge on the power grid,but the removal of a certain edge or a certain number of edges in the network lead to a substantial decline in network robustness.With the increase of the tolerance parameter,the robustness of the network is enhanced and stabilized.
-
Computer Engineering. 2018, 44(1): 299-305.Instant messaging applications are characterized by the immediacy of the messages,and their internal heartbeat mechanism drastically consumes the resources of the mobile system due to their permanent timing operations,which is especially evident on Android systems.Therefore,this paper takes Android application system as the research object,the domestic mainstream instant chat software to capture test,understand and analyze its internal implementation process,detection of mobile phone consumption of resources,then uses the Xposed framework to adjust and modify the heartbeat interface,developes the application of energy-saving instant communication application (iHeart).Experimental results show that the iHeart application can intercept the heartbeat operation while ensuring the normal sending and receiving of various application messages,and can adjust multiple timing settings compromise to reduce the traffic and power consumption.
-
Computer Engineering. 2018, 44(1): 306-310.Aiming at the weak ability of predicting the prosodic level based on the characteristics of shallow grammar and dependence-based syntactic single feature,this paper proposes an improved Chinese prosodic prediction method.By extracting the dependency syntactic single feature from the dependency parsing results of the input text and integrating key features,the dependent information fusion feature is obtained.The syntactic single feature is compared with the fusion feature through the prosodic hierarchical predictive experiment.The optimal dependence feature combination is combined with the shallow grammar feature,and the C4.5 algorithm is used to predict the prosodic hierarchy.After a lot of corpus training and testing,results show that the prosodic predictions of interdependent information fusion features are better than the predicative ones based on dependency syntax.Compared with rhyme hierarchical law prediction based on shallow grammar features,the accuracy of prosodic word is increased by 5.8%,and the precision of prosodic phrase is increased by 15.4%.
-
Computer Engineering. 2018, 44(1): 311-316.In view of the complex background and real-time requirement of the dynamic target detection in motor vehicle Assisted Driving System(ADS),a dynamic target detection method is proposed based on improved Scale-Invariant Feature Transform(SIFT) operator and improved hypothesis testing method.Firstly,the wavelet multi-resolution analysis and the position estimation of feature points between adjacent frames are used to improve the SIFT method,so as to achieve fast parameter estimation for global background motion compensation.And then,by improving the background variance estimation in the three frame difference image,the targets closing to the edge of the image which are missed in the traditional hypothesis testing method can be detected exactly.Experimental results demonstrate that,not only the superior performance of the SIFT operator is maintained and the accuracy of parameter estimation is improved,but also the speeds of the feature points matching and targets detection are improved.The proposed method can meet the real-time requirements of the system.
-
Computer Engineering. 2018, 44(1): 317-321.In aircraft cockpit noisy environment,it is hard to detect pilot speech endpoint.Aiming at this problem,a robust Speech Endpoint Detection(SED) method is proposed.The initial speech denoising is carried out by using Optimally Modified Log-Spectral Amplitude Estimator(OM-LSA) speech enhancement algorithm,and Teager Energy Operator(TEO) is used to further filter out the residual noise.Meanwhile,the ratio of the short-term energy of the speech to the spectral entropy of the sub-band is used as the parameter of the double-threshold decision,so as to detect the start and end point of the pilot speech.Experimental results show that,compared with the SED method based on energy parameter or spectral entropy parameter,the proposed method can improve the detection accuracy effectively.
2024
Volume 50 No 4
Volume 50 No 4
(Started in 1975, Monthly)
Sponsored by：
East-China Institute of Computer Technology Shanghai Computer Society
Edited & Published by：
Editorial Board of Computer Engineering
Chief Editor：Jiang Bo
Executive Chief Editor:
Hu Bing
Tel：021-67092217
E-mail: ecice06@ecict.com.cn
CN: 31-1289/TP
ISSN: 1000-3428
Sponsored by：
East-China Institute of Computer Technology Shanghai Computer Society
Edited & Published by：
Editorial Board of Computer Engineering
Chief Editor：Jiang Bo
Executive Chief Editor:
Hu Bing
Tel：021-67092217
E-mail: ecice06@ecict.com.cn
CN: 31-1289/TP
ISSN: 1000-3428