Showing papers in "Computer Simulation in 2011"

Maximum Entropy Method of Image Segmentation Based on Genetic Algorithm

Abstract: The traditional entropy threshold has shortcomings of theory and computational complexity,resulting in time-consuming in image segmentation and low efficiency.In order to improve the efficiency and accuracy of image segmentation,an image segmentation method is proposed,which combines the improved genetic algorithm with maximum entropy algorithm.First,the two-dimensional histogram based on the image gray value information is used to extract features,then three genetic operations of selecting,crossover and mutation are used to search for the optimal threshold for image segmentation.Simulation results show that the improved algorithm,compared with the traditional maximum entropy image segmentation algorithm,increases segmentation efficiency,and the accuracy of image segmentation has greatly improved,which speeds up the segmentation speed.

Research on Chaotic Feature Extraction of Ship Radiated Noise

Abstract: An improved subspace noise reduction method is proposed,for the conventional subspace noise reduction method is merely adapted to multi-signals,and the phase space reconstruction is combined to transform the single signal to multi-signal.An advanced multiple iterative noise reduction algorithm is presented,which takes the defined SNR improvement as an iterative controller.K-L transform is cited to reduce the feature dimension.This paper forms a whole ship radiated noise feature extraction system based on subspace noise reduction theory,phase space reconstruction and K-L transform.Simulation results show that the method has good effect in classifying three types of ship.

Formation Control and Obstacle Avoidance for Multiple Mobile Robots

Abstract: In this paper,the problem of formation control and obstacle avoidance for multiple mobile robots is discussedFirst the formation construction is established,and the formulation of formation control for avoiding the obstacles is givenBy using potential function,a formation control scheme is proposed for multiple mobile robots with obstacle avoidanceUnder the proposed distributed control scheme,the closed-loop stability and safety are guaranteed theoreticallySimulation results demonstrate the effectiveness of the proposed control strategies

Research on Invulnerability of Combat Net Model Based on Complex Networks

Abstract: The net model of combat was constructed and its invulnerability was studied by complex networksRobust and frangibility of the combat mode were analyzed with the change of operation loop quantityBy simulating the damage to different egdes,quantity connection between egdes and operation loops was proposed quantificationally,and this shows that cooperation between decision-making centers acts significant in NCW

Forecast of Railway Passenger Traffic Based on a Grey Linear Regression Combined Model

Abstract: The forecast of railway passenger traffic is an important foundation as well as one of the main bases for the work of railway passenger transportation.The accurate forecast of passenger traffic is an important guarantee for market-oriented railway transport enterprises to win the future.The railway transportation system is a gray system with incomplete information.This grey linear regression combined model is used to forecast the railway passenger traffic of Henan Province for the next five years.It makes up some deficiencies of the original linear regression model which has the defect of exponential growth trends and of GM(1,1) model which is lack of linear regression.Compared with single model,its forecast results are more accurate and more reliable,which not only shows high practical value,but also provides a new way for the study of railway passenger forecast.

Image Matching Method Based on Improved SIFT Algorithm

Abstract: To improve the resolution of different scales in different image matching accuracy and efficiency,this paper introduces an improved method based on scale invariant feature transform(SIFT) algorithm for image matching.SIFT feature matching algorithm is currently a hot research field,because it has good invariance of scale,rotation,illumination,based on analyzing the generating process of SIFT feature vectors,the quasi-Euclidean distance instead of Europe is as the similarity measure of feature descriptors to improve the SIFT feature matching efficiency.Experimental results show that under the condition of keeping the image matching rate and algorithm robust,the method can not only shorten the matching time,but also improve the matching accuracy.

Underwater Target Recognition Based on Wavelet Packet and Principal Component Analysis

Abstract: Underwater target classification and recognition is a research challenge of signal processing application,and as a result of complicated target signal and diverse ingredient,the character data are great and with high dimension,which needs huge calculating cost.Under this situation,a new approach to extracting noise radiated from underwater target based on wavelet packet and principal component analysis is presented.Firstly Initial characteristics are obtained from underwater target by using decomposition and reconstruction of wavelet packet.Then principal component analysis is used to get the final characteristics.The final characteristics are used by designed neural network to recognize the noise radiated from underwater target.Experiment results show that the method of extracting features has better classification effect with low calculating cost.

Development of Decision Support System

Abstract: Although the original Decision Support System concept has proposed for not more than 40 years,it has make progress in different fields by means of integrating with different subjects.The evolution of DSS technologies was discussed first,and then some technologies related to DSS were proposed,including data warehouse,OLAP,data mining,simulation and Web-based DSS.The state of widely-used Intelligent Decision Support System was described,including IDSS based on Expert System,IDSS based on Agent and IDSS based on Neural Networks.Finally,some implications for the future of the field were discussed.

Research on Intrusion Detection System Neural Networks and Principal Component Analysis

Abstract: Research on the network security question of intrusion detection.According to the features of high dimensional,nonlinear and redundant of the network intrusion data,and the problems that the it is difficult to reduce the dimension and the testing rate is low in traditional methods,An intrusion detecting method is proposed based on the analysis of the main genetic neural network methods.First,the dimension of network intrusion data is reduced using the principal component analysis to eliminate the redundant information and simplify the neural network's inputs.Then,using genetic algorithm of neural network weights,the learning speed of neural network is accelerated.Finally,the neural network model is adopted to optimize the data after the principal component analysis and draw the nonlinear rule of network intrusion detection data.Through the network intrusion KDD CUP 99 algorithm for data collection verification experiment,experimental results showed that,compared with other network intrusion detection methods,this method is fast in learning,and has high detection accuracy and low fail and error rate.It is a kind of efficient and real good network intrusion detection method.

Load-Capacity Model for Cascading Failures of Complex Networks

Abstract: To research on processes and features of cascading failure on complex networks is helpful to guide system construction and improve its robustnessBased on related works,a load-capacity non-linear model was proposed,which is more suitable for real networkThe simulation was executed on B-A scale-free network and Internet AS level topologyIts results show that the model is feasible and negative exponential relationship is found between model parameters under certain conditionsConsidering network cost and robustness,it is validated that the model can defense cascading failure better and investment costs are smaller in case of obtaining higher robustness

Study on Simulation of Multiple-receiver Synthetic Aperture Sonar Imagery Based on Wide Swath

Abstract: Multiple-receiver synthetic aperture sonar(SAS) technique has been adopted to remove image ambiguities,in many papers,but the image distortion caused by the stop-and-hop approximation is usually ignored or underestimated.A robust multiple-receiver SAS processing method is proposed to process return data,taking example for Range-Doppler algorithm,after analyzing the system phase error in SAS with multiple-receiver.The Fresnel approximation is discarded and the phase error introduced by the stop-and-hop approximation is corrected exactly,so this method can be used for imaging in SAS with low center frequency and wide swath.The simulation results show the validity of this method.

Research and Development of Virtual Maintenance Simulation Technology

Abstract: It is virtual maintenance simulation technology that has been widely used in the three fields of maintainability design,supportability planning,and maintenance personnel training and achieved great economic and social benefits.Therefore,it becomes the current research hotspot.So numerous cases of virtual maintenance simulation at home and abroad were analyzed and summarized.In view of the extension of simulation object and the application of new technology,the new definition of virtual maintenance simulation technology was proposed.Realization flow of the virtual maintenance simulation technology was summarized and a series of involved key technological problems were investigated.Many related popular software realization platforms were analyzed and compared.Finally,the prospect of the future work in this area was also presented.

Wireless Sensor Network Deployment Based on Genetic Algorithm and Simulated Annealing Algorithm

Abstract: Wireless sensor nodes deployment optimization problem is studied.Wireless sensor network deployment determines its capability and lifetime.Traditional genetic algorithm generates local optimal problem easily in wireless sensor nodes deployment optimization process because the probabilities of crossover and mutation are fixed,thus the results are not ideal,and network lifetime is too short.In order to optimize network deployment and improve the network life,this paper puts forward a wireless sensor nodes deployment optimization method based on genetic algorithm and simulated annealing algorithm.In this method,sensor nodes deployment optimization is transformed into combinatorial optimization problem,and network nodes are expressed as a grid,using genetic algorithm to search the optimal deployment,and simulated annealing algorithm is used to the population change,for the search speed improvement.Simulation results show that the proposed algorithm,compared with the simple genetic algorithm and simulated annealing algorithm,has higher efficiency in deployment and more alive nodes in network,which prolongs the network lifetime.

Research on Path Planning for Mobile Robot Based on Improved Genetic Algorithm

Abstract: Premature and lower convergent speed is two puzzling problems in applying genetic algorithm,a genetically simulated annealing algorithm of optimum path planning for mobile robots is proposedChanging of two-dimensional codes into one-dimensional codes is adopted to simplify the encoding pathAn initialization population was produced based on genetic algorithm,and the fitness value of each path is evaluatedAn efficient temperature updating function was devised through a series crossover and mutationAnd by adopting the random moving rule of Metropolis algorithm,a global optimal path was obtained from the starting point to the target pointFinally,the feasibility and efficiency of this algorithm are verified in the Matlab environmenThe simulation results demonstrate that the proposed algorithm has achieved considerable improvements in convergence speed,search quality and the best path compared to the basic genetic algorithm

Research of PreciseLaser Spot Center Location Algorithm

Abstract: Traditional sub-pixel location algorithm often has some problems such as poor anti interference capability,low location precision and complexity of software realization.In order to meet the requirements of precise laser spot center positioning in distortion measurement system,the curve fitting sub-pixel location algorithm of barycenter is proposed.On the basis of barycenter method,curve fitting algorithm is proposed to improve the precision of laser spot center location.As the image preprocessing is added,the algorithm effectively reduced noise and enhanced noise robust.Only few data points are calculated by the algorithm which not only has simple language description,but also effectively saves system resources.The simulation results indicate that the sub-pixel location algorithm is a practical precise location method and meets the accuracy requirements of micro distortion measurement.

Path Planning for UAV Based on Mixed Ant Colony Algorithm

Abstract: The key and difficult problem of UAV path planning is how to satisfy safety and real-time environment,meanwhile,a global path-planning and a local path-planning are considered to improve operational efficiency and survival probability.For this question,according to the existing research of UAV path planning,the method of synthesizing Ant Colony Algorithm(ACA) and Artificial Potential Field(APF) was discussed.ACA was used as a global route-planning algorithm,and APF was used as a local route-planning algorithm.Simulation results verify that the efficiency of the algorithm can provide some reference value to related researchers.

Task Assignment Scheme Based on Contract Net

Abstract: The Multi-UAV(uninhabited aerial vehicles) cooperative target assignment is an important part of Multi-UAV cooperative control.For multi-UAV cooperative reconnaissance problem,the targets are allocated in order to improve operational efficiency,reduce costs and Task completion time.The mathematical model for Multi-UAV task assignment problem was build firstly,based on the initial allocation,a target assignment method based on contract-net was presented,load factor was introduced,finally,a rational allocation of tasks was achieved by iterative implementation of the sale contracts and swap contracts.Simulation results demonstrated that the assign schemes presented by the algorithm have good optimal effect and good scalability.

Research on Precipitation Prediction Based on Time Series Model

Abstract: The forecasting accuracy should be improved in the study on precipitation prediction.It is difficult to predict climate because of the dynamic characteristics of sample set as well as the effect of environmental factors.In order to improve the accuracy,a novel model based on time series and environmental factors was introduced in this paper.Firstly,the environmental factors were nonlinearly screened by support vector machine(SVM).Secondly,the order was estimated by controlled autoregressive(CAR).Lastly,reliability of SVM-CAR was validated by one-step prediction method.The simulation result of precipitation forecasting showed that this method has the advantages of high-precision and good prospect in drought and flood forecasting.

Simulation Research on Mobile Robot Path Planning Based on Ant Colony Optimization

Abstract: Ant colony algorithm performs better than traditional optimal algorithms in many fields.In order to improve the efficiency of mobile robot path planning,a mobile robot optimal path planning method is proposed based on ant colony algorithm.This method firstly build up the environment modeling by grid,the starting point is used as ant nest while the ultimate goal point position as a food source.The optimal path is found by ants cooperation.Back strategy is used to prevent robot path search fall into deadlock state.Experimental results show that this algorithm can improve the efficiency of the robot path planning,get the optimal robot path and avoid the obstacle safety.

Stock Price Prediction Based on PCA and BP Neural Network

Abstract: In stocks decision-making study,there are lots of stock prices factors and there are highly nonlinear and redundant information among factors,the traditional stock prediction method cannot eliminate data redundancy and capture nonlinear feature,and prediction accuracy is lower.In order to improve the prediction accuracy of stock prices,a stock prediction model(PCA-BPNN) is put forward based on principal component analysis(PCA) and BP neural network(BPNN).Firstly,PCA is used for selecting variables to reduce the complexity of BPNN model,eliminating redundancy among the factors,reducing the input data,and improving the speed of BPNN.By using BPNN,the model was built with the reserved factors.The PCA-BPNN is test with stock 600008,and the results showed that PCA-BPNN can improve the prediction accuracy and is a high efficient and accurate stock prediction model.

A Navigation Method and Its Simulation for UAV Formation Flight

Abstract: High-precision location for all members is the core technological issue for UAV formation flight.A navigation algorithm base on INS/GPS integrated navigation and relative measurements was put forward to overcome this issue.INS errors of all members were estimated with Kalman flitering and precise navigation information was obtained after correction.Simulation results verify the validity of this algorithm and show high geodetic navigation accuracy.

Investigation of Direct Adaptive Controller for Tank Gun Elevation Control System

Abstract: A direct adaptive control method of tank gun elevation control system was put forward using the CGT-based direct adaptive control principles.The plant was divided into strict positive real(SPR) part and the other part by its structure.A CGT-based direct adaptive controller and a robust controller were designed for the two parts respectively to control the gun vertically.The method has no need to identify the parameters of the plant,so the structure of the system is simple.It also has a strong adaptive ability to variations of the parameters.Simulation shows that the method is correct and feasible.

Approach to Information Systems Security Risk Assessment Based on Fuzzy-BP Neural Network

Abstract: In order to improve the accuracy of the information security evaluation and reduce the effectiveness of expert's subjectivity,this paper put forward an improved the method of fuzzy theory and BP neural network model for evaluation.Corresponding evaluation set membership matrix was created through analysis of assets value.The information security risk levels were obtained from learning of fuzzy theory and the BP neural network.Finally,simulation experiments show that this model can provide precise result of evaluation information system risk assessment.Compared with the models,the accuracy is enhanced,and it is an effective information safety risk assessment method.

Dynamic Modeling and Characteristic Simulation of Charging-discharging System for Electronically Controlled Air Suspension

Abstract: In order to obtain the dynamic characteristic of charging-discharging process for Electronically Controlled Air Suspension(ECAS),the charging-discharging dynamic model of ECAS was created by combining the thermodynamics process of variable mass system and vehicle kinetic equationConsidering the nonlinearity of charging-discharging process,the characteristic simulation was carried on by building the Matlab/Simulink modelIt has been shown that the charging-discharging process must subdivide two parts of the variable mass opening system and the closed balancing system after closing the electromagnetic valve,and that the latter makes "over-charging" or "over-discharging" when the ride height switchingThe effecting rules of the charging-discharging process are achieved by simulating analysis under the different charging-discharging time and suspension parameters(egmass,damping) changing,and the simulating results provide a foundation for the height control strategy of air suspension

Job-Shop Scheduling Based on Genetic Algorithm and Simulated Annealing Algorithm

Abstract: During the iterative process of standard genetic algorithm(GA),the premature convergence of population decreases the algorithm's searching ability.Through analyzing the reason of population premature convergence during the renewal process,by introducing the selection strategy based on simulated annealing algorithm(SA),a genetic algorithm based SA model was proposed,and its application to job-shop scheduling problem(JSP) was given.The selection strategy based on SA keeps the population diversity during the iterative process,thus overcomes the defect of premature convergence.Simulation for 30 benchmarks of JSP indicates that compared with GA algorithms,SA-GA gets better result within shorter period,thus certifies the improvement of the algorithm's searching ability by annealing mechanism.

Study on Contour Force Rendering Technology Oriented to Image Perception

Abstract: With the development of the haptic interaction technology,the force perception will play an important role in more fields,such as in entertainment,blind help,and science popularityForce rendering is a key problem to realize the realistic force perceptionThe image-based 2D contour force rendering method was investigatedThe user interacted with the planar images by operating the 2D force feedback mouseWith the force feedback,the user could perceive the contour shape of the imageThe target image boundary was firstly extracted and the discrete pixel points were fitted to a polygonal contour curveThen the collision detection was done according to the relationship between the positions of the virtual tool and the contour as well as the minimum distance from the virtual tool point to the contourBased on the continuity of the contour force,the contour force rendering algorithm was proposed which ensured the system stableAn experiment was conducted to identify the animal sign by the fedback contour forceThe performance,including real time response,stability and recognition rate was analyzed which verified the feasibility and the effectiveness of the presented method

Study on Feature Extraction of Ship Radiated Noise Based on Higher Order Spectrum and Cepstrum

Abstract: The feature extraction of ship radiated noise is a research challenge in signal processing applicationDue to the complexity of marine environment and the special nature of underwater acoustic channel,it is difficult to obtain ship radiated noise' signals from the complex background noiseUnder this situation,this paper proposed the several extraction methods of higher order spectrum and cepstrumFirstly,it selected the appropriate extraction method based on different feature of ship radiated noise,in order to extract the feature signals which has more line-spectrum and the ability to suppress interference noiseAt last,real signals as well as the simulated signals were extracted by the characteristics of the bispectrum,three spectrum,11/2 spectrum of higher-order statistics and cepstrumWe also pointed out the advantages of using the extraction methods to extract ship radiated noisesThe results of the experiment show that extraction methods are feasible,which are helpful to provide effective feature for target recognition

Human Behavior Recognition Based on HMM

Abstract: Hidden Markov Model is a statistical analysis model,with great abilities to analyse time variant patterns.In order to increase the effectiveness of image information and improve the capability of recognizing the image sequences of human behaviors,a novel method based on the combination of Fourier descriptor and Hidden Markov Model is proposed to recognize the serial image of human actions.The paper first obtains the binary image sequences,extracts the Fourier features with the characteristics of rotation,translation and scale invariances,and proposes a kind of Fourier transform based on the center distance,then uses the improved Hidden Markov Model to classify the extracted feature vectors and gets the recognition results.The experiment results show that the method proposed in this paper is effective and feasible,and the recognition rate is related to the state number and observation number of HMM,and the appropriate HMM classifier can make the recognition rate of the system rise be above 90%.

Blade Flutter Analysis Based on Fluid Solid Coupling

Abstract: Blade flutter can cause serious damage,and it is a part of the study of fluid solid coupling.In order to predict and analyze blade flutter,exhaustive analysis of fluid solid coupling and the process of numeric simulation are presented here.And fluid solid coupling simulation is accomplished by using ANSYS and CFX in numerical calculation of structure and flow field,with a coupling bench which can transfer fluid pressure and structure displacement.Under different inlet velocities and attack angles,blade and fluid coupling is simulated.And blade displacement response is acquired to determine whether flutter will occur or not.Simulation results indicate that flutter frequency is close to natural frequency of the blade,inlet velocity and attack angle are important factors which influence the aero-elasticity stability.

Application of Improved K-means Algorithm in Telecom Customer Segmentation

Abstract: In order to implement differentiation of telecommunications customer marketing and customer service strategies,customers need to be accurately classified.K-mean algorithm is an important method for telecom customer classification in the data mining technology,but in actual classification process,k-means clustering algorithm is sensitive to initial value and easy to fall into local optimal.In order to improve the telecom customers' classification accuracy,a new clustering algorithm is put forward to improve the classification accuracy.The improved k-means clustering algorithm's initial clustering K and clustering center are adaptively determined by hybridization,operator selecting and classification feature.Simulation experiments show that the improved k-means clustering algorithm can solve the problem that traditional k-means clustering algorithm easily trapps into local optimal,improve customer classification accuracy and reduce errors greatly.