Based on robust L₂-L_∞/H_∞ performance criteria for continuous time system, a new approach to design robust L₂-L_∞/H_∞ controller was proposed, which was then transformed into convex optimal problem with constrains of linear matrix inequalities with low conservative. Using this method, the vehicle chassis integrated control strategy was developed to coordinate active front wheel steering system and active suspension system. Typical cases were simulated using Matlab/Simulink and Carsim. Results show that the robust L₂-L_∞/H_∞ controller for vehicle chassis integrated system has good control effect, and can significantly improve the stability of the travelling direction and anti-rollover capability of the vehicle.

In order to improve the vehicle stability under complex working condition, an electronic stability control system based on hierarchical strategy is proposed. The up layer of the controller is a fuzzy Proportional-integral-derivative (PID) controller, and the controlling variables are the yaw rate and the side-slip angle. The desired yaw moment is calculated by the upper layer. The target slip rate is calculated based on the relationship between the yaw moment and the slip rate. Then, the lower layer is applied to control the slip rate. Finally, the ESC algorithm is verified by Matlab/Simulink. The winter road test shows that the proposed control strategy can improve vehicle stability performance.

A method to optimize the macro-parameters of helical gears in transmissions is developed to reduce meshing vibration in initial design phase of the transmission gearbox. In the optimization model, the optimization variables include the modulus, the helical angle, the tooth wide, the top gap coefficient etc. The optimization target is the synthesis of reducing meshing vibration and reducing the volume of the gear pair. The constraint conditions include the standard requirements and the strength requirement. The fluctuation value of the dynamic transmission error is considered as the index of the vibration. To calculate the dynamic transmission error of the helical gear pairs, the helical gear pair is approximated as the superposition of many smaller spur gear pairs, and the phase difference of the meshing angles of adjacent smaller spur gear pairs is considered. The deformation calculated by Ishikawa method and the circumferential deformation are considered in the dynamic transmission error and the dynamic meshing stiffness calculation of the smaller spur gear pairs. The optimization results show that the method developed in this paper can effectively reduce the transmission error of the helical gear by 41% and reduce the volume by 1.34%.

The structure and load forms and condition constraints of SUV Body-in-white (BIW) are different from that of sedans. In order to quickly realize the automatic lightweight structure of a SUV BIW, an implicit parameterization model of the SUV BIW is built. First, correlation analysis method is applied to analyze the influences of 90 pre given potions, shape, thickness and other designing variables on the body structural rigidity, and the variables with lightweight potential are selected. Then, focusing on 30 selected designing variables, a multi-performance lightweight optimization is carried out to achieve good lightweight effect. Results show that the proposed implicit parameterization model and multi-performance lightweight optimization based on correlation analysis can be applied to SUV BIW structure lightweight design and are of engineering significance.

Based on the data of collision accident between car and electric bicycle obtained by CIDAS, the cyclist injury mechanism and the causes are studied. The THUMS biomechanical dummy is used to represent the cyclist in the accident. Comparison of the simulation results with real accident data shows that the method is reliable. Analysis suggests that the probability of serious leg injury is higher than other body part injury in the collision accident. The structure of the frontier of the engine cabin is the key parts to reduce the leg injury. The simulation results can provide valuable reference for designing safety regulation and new car development.

In order to improve the accuracy of short-time traffic flow prediction, a short-term traffic flow prediction model based on singular spectrum analysis and Combined Kernel Function (CKF) Least Square Support Vector Machine (LSSVM) is proposed. Singular spectrum analysis technology is used to filter out the noise of traffic time series. Then, the processed traffic flow data are used to train the CKF-LSSVM, and the parameters of the model are determined by particle swarm optimization algorithm. Finally, validation and comparative analysis of the model are carried out using the measured in Xiamen, China. Experimental results indicate that the proposed model has good prediction performance and can effectively improve the accuracy of short-time traffic flow prediction.

Existing shortest path searching algorithms do not consider the coordinate control of arterial intersections, which leads to unsatisfactory searching results. To overcome this drawback an urban shortest path algorithm is developed, which takes the coordinate control of arterial intersections into consideration. First, with the analysis of the coordinate signal timing at the arterial intersections and the offset parameter, an improved signal intersection waiting time model is proposed. Second, with the combination of IDA^* algorithm and the improved signal waiting time model, an improved IDA^* algorithm considering the coordinate control of arterial intersections (AICIDA^*) is developed. The AICIDA^* algorithm is compared with two traditional algorithms and an improved A^* algorithm considering the coordinate control of arterial intersections (AICA^*) through a case study, in which the path time cost and computing time are taken as the evaluation indexes. Results show that the proposed AICIDA^* algorithm can obtain the better path time cost in a shorter computing time.

In order to reduce the bus delays in a bus stop, which located near signalized intersection, the bus stops are classified into two categories according to their locations on the road. A bus delay model is established based on the shockwave theory by analysis of the interaction influence between bus stop and traffic flow. If the bus stop is located upstream the intersection, bus delays include three cases and the expected delay values are obtained by probabilities and corresponding delay case. If the bus stop is located downstream the intersection, the coordinate system is established to describe travelling track of both buses and cars, and the bus delay model is constructed. Simulation was carried out with aid of Matlab. Simulation results show that, with the growth of traffic flow, signal cycle and dwell time, the bus delay increases if the bus stop is located upstream the intersection. If the bust stop resides downstream the intersection, the bus delay decreases first and then increases as it approach the upstream stop.

Considering the real demand for pedestrian diagonal crossing at signal intersection, the operation effects of different signal phase settings and pedestrian crossing organization schemes were analyzed using the time-space diagram. Then, a design approach of pedestrian green wave for continuously crossing signalized intersection was proposed. According to the pedestrian arrival and departure time, the delay model of pedestrian continuous crossing along the clockwise, counterclockwise and unfixed directions were deduced. Finally, the general delay model for pedestrian continuous crossing was established. The results of case study and simulation show that the general delay model can predict the delay of pedestrian diagonal crossing accurately. This model can also help to acquire the best signal phase sequence and pedestrian crossing organization scheme. Therefore, the green wave can be obtained and the delay of the pedestrian diagonal crossing can be reduced.

A rout planning and coordinated scheduling model is developed to minimize the passenger's costs, including the waiting time and the difference between actual transfer time and the expected transfer time, and the operator cost, the total travel time of the vehicles, with considering the known passenger demands, fleet size and the departure time of the main line at the transfer hub. As the objective of this model is a nonlinear function, the model is transferred into a mixed integer programming model for the convenience of resolution. In the resolving process, a gravity-based heuristic method is employed to solve the model rapidly. A case study is given to illustrate the reliability and practicability of the proposed model.

The difference between lane change intention stage and lane keeping stage was studied by analyzing the driver lane change performance and the vehicle movement parameters. Based on BP neural networks model and D-S evidence theory model, the real-time identification of lane change intension tests were carried out. The results show that that sample identification accuracies of the two models in three seconds before the lane change are 78.3% and 45.2% respectively. The identification accuracies just at the moment of lane change are 99.13% and 86.96% respectively. By verifying the two models with random sample data, the identification accuracies for lane change intension are 86% and 96%, while for lane keeping are 21.05% and 78.95% respectively. The maximum time of the models for accurate identification is less than 0.5 second. It indicates that the evidence theory identification model is more applicable and reliable in both accuracy and timeliness.

A road grade estimation method, which uses road grade change rate information, was proposed. The change rate of road grade is often ignored by researchers while using modern control theory to solve on-line road grade estimation, which may cause time lag inevitably. To eliminate this shortcoming, the acceleration measured from longitudinal sensor was introduced to express the change rate of road grade. Then, the state transition matrix including vehicle speed, vehicle acceleration and road grade was built. Finally a Kalman filter for road grade estimation was constructed. Experimental results show that the propose road grade estimation method considering the grade change rate information can reduce the time lag obviously.

In order to effectively identify risk status while driving, a driving risk status identification model is proposed based on the information of driver operation and vehicle status. According to the rate of electrocardiogram (ECG) and traffic violation behaviors, the driving risk status is classified into four levels. Using Markov blanket algorithm, five factors are selected as the feature set, including the speed, the longitudinal acceleration, the rate of front wheel angle, the vehicle position and lane departure. Then, the algorithm of Hidden Naive Bayes (HNB) is employed to establish the driving risk status identification model. The results show that the accuracy of HNB is 90.6%, increasing 14.1%, 13.9 and 13.0% compared with Bayesian Network (BN), Naive Bayes (NB) and Radial Basis Function Neural Network, respectively. The results of ROC curve indicates that the model presents high predictive power. The conclusions can provide theoretical support for designing dangerous driving status recognition equipment based on vehicle and driver operation information.

Traditional damage identification methods are difficult to accurately identify the damage location and extent of irregular-shaped bridge because of the complex structure of the bridge. To overcome the shortcoming of these damage identification methods and the uncertainty of factor influence, the factor effect on damage identification indicator is synthetically analyzed, and a novel damage identification approach for irregular-shaped bridge is proposed based on modal flexibility difference curvature and Support Vector Machine optimized by Genetic Algorithm (SVM-GA). First, the damage identification effect on the indicators (modal shape change, modal curvature difference, modal flexibility difference and curvature) is studied under the impacts of damage severity, number of sensors and modal orders. Second, the two stage damage identification method for irregular-shaped bridge is presented. On the basis of determining the damage location using modal flexibility difference curvature, the damage extent is forecasted and identified based on SVM-GA. Damage identification results demonstrate that the proposed method is feasible and accurate.

The fatigue failure mode and anti-fatigue performance of Carbon Fiber-reinforced Polymer (CFRP) post-tensioned beams are systematically studied by fatigue test. The results show that, when fatigue failure occurs, there are no fractures on the steel bars and the external prestressed tendons, but the upper concrete is crushed. With the increase in the number of fatigue loading times, as the fatigue loading level S<0.43, the growth of CFRP tendon strain does exist but is not obvious; however, as S>0.43, CFRP tendon strain increases significantly. After certain number of fatigue loading times, the relationships between the steel strain, the concrete strain and the loading are nonlinear, indicating that the beams are damaged due to fatigue; while the relationship between the strain of CFRP tendons and the loading is linear. Under the same static loading, the higher the number of fatigue loading times is, the larger the strains of the steel bars, the concrete and the CFRP tendons are. Compared with the common steel reinforced concrete beams, when S=0.48 the fatigue life of the CFRP post-tensioned beam is increased by 30%.

Based on theory of elastic mechanics, the bonding stress model of adhesively bonded single lap joint under eccentric load was established. The effects of different parameters on the interfacial shear stress were investigated. A quantitative design formula of the single lap joint was given. The results show that the maximum shear stress is located at the bond corner, which will be first damaged. However, the bond failure can be avoided by increasing the width and length of the single lap joint. The shear stress decreases nonlinearly with the increase in distance. Besides, if the hardness of the overlap plate is lower, the bond stress at the corner will be higher under eccentric load. The results of this study provide a reference for engineering practice.

The structure, main features and functions of active thermo-atmosphere combustor are described. Then, the main topics and achievements in the research of turbulent spray auto-ignition and flame stabilization are reviewed. The studies on this type of active thermo-atmosphere combustors are classified according to the fuel characteristics. The research advantages and features of controllable active thermo-atmosphere combustor are analyzed. The directions for future research of such combustor are predicted.

The 3D model of a Diesel Particulate Filter (DPF) with asymmetrical channels was established in Fluent. Based on numerical model, the fluid flow and pressure drop characteristics were simulated, the soot loading process was analyzed, and the pressure drops in asymmetrical channels and symmetrical channels were compared. The results show that, along the inlet channels, the flow velocity magnitude increases at first and then decreases, while the static pressure decreases first and then increases. Along the outlet channels the flow velocity increases while the static pressure decreases. At the beginning of soot loading process, larger amount of soot loading is found in the downstream half of the channel. As the soot loading increase, the distribution of soot along the axial direction becomes uniform. The pressure drop in the asymmetrical channels is lower than that in the symmetrical channels. The coefficient of the absolute soot loading of the asymmetrical channels is above 0.3. As the inlet-to-outlet ratio increases, the influence of the inlet mass flow rate on the pressure drop becomes more significant.

For cone valve throttle orifices with different structure parameters, first, the characteristics of triangular (45°, 53° and 60°), U-shape and rectangular throttling notches are analyzed and the corresponding calculation formulas of the valve port areas are deduced. Then, the flow coefficients of different valve cores in cone valve are obtained according to the experiments on the flow and pressure characteristics and the Matlab program. Results show that the flow coefficients of the cone valve are not only related to the system pressure and valve port area, but also related to the configuration feature. The flow coefficient of cone valve increases with the pressure, and it decreases with the increase in the orifice area. The stability of the flow coefficients of triangular valves with 53° and 60° throttle grooves is better than the other three kinds of structure; and the flow coefficients of triangular valve with 45° throttle groove is most unstable.

In order to make full use of the energy saving ability of hydraulic common pressure rail system and enable the driving system have four-quadrant operating characteristics, an adaptive reverse driving system based on hydraulic transformer is presented. By controlling the hydraulically operated check valve with pressure cross feedback, the driving system can realize adaptive reversion by changing the control angle of the hydraulic transformer. The system model is built and the operating characteristics are analyzed. The results indicate that the system response is fast when the system switches from driving mode to braking mode, the speed of hydraulic transformer decreases sharply. After the braking pressure built up, the hydraulic transformer operates in the steady condition and the speed decreases gradually. Under the braking condition, the system can realize regenerative braking and save part of the kinetic energy. The energy is stored in the hydraulic accumulator.

A flow-control energy-saving rotary system was proposed to reduce the energy consumption caused by the mismatching between the output flow of the pump and the demand flow of the rotary motor during the rotary starting process of hydraulic excavator. The overflow signal of the overload buffer valve was converted to a pressure signal using a flow control valve, and the pump displacement was adjusted according to the pressure signal. Simulation and experiment results show that the flow-control energy-saving rotary system can effectively reduce the energy consumption during the rotary starting process that over 20% energy can be saved in comparison with the traditional rotary system. The overflow has long been existent, which meets the torque demand during the rotary starting process, and the time of the starting process stays the same with that of the traditional rotary system.

In order to accurately solve the kinematic parameters and mechanical parameters of articulated tracked vehicle in turning process, the kinematic model and mechanical model of the vehicle's steering hydraulic cylinder, and the kinematic model and mechanical model between the track and soil are established, in which the soil shear deformation and crawler slip are taken into consideration. The formulas of the angular velocity, angular acceleration of steering hydraulic cylinder, the steering radius, the shear resistance and steering resisting moment are theoretically derived. These formulas are solved numerically for a specific articulated tracked vehicle as an example. Finally a virtual prototype model of the articulated tracked vehicle is adopted to verify the theoretical model. This research can provide theoretical basis for the smooth steering control and structural optimization of articulated tracked vehicle.

A new model of involute spur gear of form inking roller of offset press that contains rotational and axial velocities is proposed to study the lubrication behaviour. The model is based on the finite line-contact mixed Elastohydrodynamic Lubrication (EHL) formulas, which is capable of handling the numerical simulation gear lubrication with different rough surface patterns. On the basis of geometric analysis, the governing equations are built up with the help of the mixed EHL theory. Then, the efficient multigrid method is applied to solve the governing equations after the nondimensionalization and discretization. As results, the numerical simulations can accurately predict the film pressure distribution, the film thickness distribution, the compositive friction coefficient and the asperity load ratio. It is shown that the axial velocity of spur gear intensifies the stress concentration on the gear tip. Different rough surface patterns obviously influence the lubrication behaviour of the spur gear. The film pressure of longitudinally oriented surface reaches a maximum of 593.1 MPa under the condition of non-axial velocity. The rise of the film pressure of longitudinally oriented surface reaches a minimum of 188.7 MPa only under the condition that the axial velocity is not zero. It is also shown that both the axial velocity and the rough surface pattern obviously influence the friction coefficient.

An integrated design method of aerodynamic and structural performance for wind turbine airfoil is proposed based on conformal transformation theory. In this method, the parameterized expression of an airfoil is presented by combing Theodorsen theory and B-spine curve. In order to optimize the aerodynamic and structural performance, an optimization mathematical model of the airfoil is established. The RFOIL software is used to solve the aerodynamic performance and a MATLAB code is used to solve the structural performance. The airfoil optimization is accomplished using an improved genetic algorithm. A new airfoil named CQUL210 with 21% maximum relative thickness is designed. Compared with the worldwide well known wind turbine airfoil DU93-W-210, of which the maximum relative thickness is also 21%, CQUL210 has higher aerodynamic performance both in smooth and rough transitions in the range of the design attack angle. Finite element analysis indicates that CQUL210 has higher structural performance than DU93-W-210. The method of this study can be applied to increase the wind capturing power and reduce the mass of the turbine blade.

An analysis method of replacement time of system components of NC machine tools based on time-correlation is proposed. The aims are for making reasonable maintenance strategies and accurately estimating the stock of spare parts. First, taking the system components of NC machine tools as the research object, the trend and stability of the fault data are tested. Then, considering the correlation between the fault times of the components, the fault time sequence of the system components is modified by Johnson method. The failure process model of the system components is established through parameter estimation and hypothesis testing. Assuming that the failure rate of the test machine tool components obeys normal distribution, the index of replacement failure rate is estimated by fitting the confidence interval of normal distribution's low limit. Finally, the equipment maintainable time and spare part quantity for certain time can be calculated. A case study verifies that the proposed analysis method is practicable and effective.

The three rollers axial knurling process of assembled camshaft is investigated by experiment and numerical simulation, including the distributions of stress and strain and their changes on the surface of the tube, the formation process of tooth profile of the tube, the effect of parameter of knurling tools on the tools' load, the assembly interference quantity, the press-fit and torsional strength. Results show that the tooth profile formed by three point knurling technology is complete and satisfactory, and the metal can flow sufficiently, and the stress does not exceed the material failure stress. The tooth height and addendum angle are important factors in determining the tool loading, assembly interference quantity, press-fit and torsional strength. The simulation and experimental results are in good agreement, which can provide a reliable reference for the selection of axial knurling process parameters.

The computation criterion of unit displacement of multi-point 3D stretch bending die was built. The fast iterative process of controlling multi-point die shape was achieved by using multi-phase compensation factors. Finite element simulation and experiment results show that the part springback value can be reduced from 6.87 mm to 1.03 mm, which demonstrates that the shape deviation and springback can be effectively controlled by the optimized die shape.

High Current Pulsed Electron Beam (HCPEB) was employed for the surface irradiation of Cu-C composite material made by high sintering, where C is in the form of flake graphite. The irradiating times were 1, 5, 20 and 40 respectively. The phase composition and microstructure of the irradiated samples were analyzed in detail using XRD, SEM and TEM. The surface friction and wear properties of the irradiated samples were also examined. Results of XRD indicate that the diffraction peak shifts to high angle after HCPEB irradiation, and the solid solution occurs between Cu-C. Furthermore, C atoms and Cu atoms forms interstitial solid solution partly, and the maximum calculated solid solubility reaches 2.24%. Microscopic results show that high density Cu and C nanoparticles are formed on the irradiated surface, and high density crystal defect is induced in the matrix Cu at the same time, which provides a large number of diffusion channels for the diffusion of C atoms. Furthermore, the friction and wear properties of the material are improved obviously after HCPEB irradiation.

The temperature uniformity of hot-stamping tool with conformal cooling channel was optimized integrally based on single-objective optimization. An optimal design of hot-stamping tool with conformal cooling channel was presented on the basis of heat flux boundary. The ratio of channel center point on the offset contour lines, the radius and the distance between channel center and tool surface are set as the optimal variables. The temperature uniformity of the tool surface is given as optimization objective. Combing with the pointer optimizer, the optimization of the shape and position parameters of the hot-stamping tool with conformal channels was carried out. The maximum temperature, the average temperature and the temperature uniformity were improved by the optimization. Moreover, the cooling performances of the tool before and after optimization by applying heat flux boundary or temperature boundary present the same trend. It is proved that the simulation approach with direct heat flux boundary is reliable and effective.

An airfoil mimicking owl wings is used to determine the noise generation mechanism of silent owl. The flow field around the bionic airfoil suggests two noise sources caused by the leading-edge separation, one is the reattached turbulent boundary and the other one is the vortex shedding detached from leading-edge bubble. The reason for owl wing noise in low Reynolds number is the interaction between the turbulent boundary layer and the sharp trailing-edge of the bionic airfoil. Furthermore, as a promising means, the passive porosity technology is applied for the trailing-edge noise reduction of the airfoil. The static pressure fields around the airfoil verify that the porous trailing-edge can alleviate pressure change by the suppression of the vortex shedding and the damping of broadband turbulent boundary layer fluctuation. The relevant noise spectrum also indicates a pronounced noise reduction potential in excess of 10 dB, but in dependence on the flow resistivity. The findings of this study may be used as reference in the design of silent aircraft.

In order to understand how elasticity works on reducing energy dissipation and improving locomotion efficiency of hopping robots, the passive hopping behaviors of hopping robot models with rigidly or elastically suspended load are examined. The influence of different system parameters on energy dissipation is discussed. Theoretical analysis and simulation results show that hopping robots with elastically suspended load exhibit lower energy dissipation than those with rigidly suspended load, because part of the kinematic energy is stored in the suspension spring during the stance phase. As the suspension damping increases, higher stiffness will be needed for the suspension spring to minimize the energy dissipation during hopping. During the flight phase, hopping robots with dissipative elastic suspension have energy dissipation which is, however, much smaller than that during the stance phase, therefore the effect of energy dissipation reduction will not be affected.

Considering that the last three wrist joints of a manipulator are highly dexterous and hove no influence on the position of the writ center point when moving, a mathematical model of the manipulator is reconstructed with origin at the wrist center point. Thus, the base of the manipulator becomes the nominal end effector with fixed orientation. All possible locations of the new end effector can be viewed as the Base's Workable Location Space (BWLS). The BWLS moves along with the wrist center point during the painting task. Through analysis of the intersection set of the BWLSs, the feasible base position area of the manipulator is established. Motion simulations on painting plane and curved surfaces referred to the concept of BWLS are performed. The results show the optional base position area of the manipulator can be quickly obtained and the once stop painting area is extended compared to the fixed rules.

A mathematical model of Permanent Magnet Synchronous Motor (PMSM) based servo system was established. A log-type barrier function was designed as control Lyapunov function to constrain the angular displacement tracking error in a certain scope. An adaptive update law was designed to estimate the matched uncertainties, and high accuracy control was implemented. Based on the vector control principle, fast terminal sliding modes were designed to ensure that the q-axis and d-axis currents reach the given values in a finite time. Applying the Lyapunov method, it was proved that the variables of the closed loop system are uniformly bounded and the tracking error can converge to a small neighborhood of the origin. The simulation and experiment results demonstrate that, compared with the traditional backstepping control, the robustness and accuracy of the proposed control method are obviously improved.

Considering actuator failures and actuator saturation constrains, the robust-tolerant control was investigated for Networked Control System (NCS) with α-safety degree. First, based on Discrete Event-triggered Communication Scheme (DETCS), a networked closed-loop failure system is built, which depends on the state time-delay and event-triggered condition. Second, in combination with Lyapunov stability theory and the three given definitions of safety-degree, the sufficient condition of invariance set with α-safety degree is derived for a networked closed-loop failure system with disturbance rejection performance, which is the time-delay/event-dependent condition. Finally, the co-design approach between fault-tolerant control and network communication is presented. Simulation is carried out and the results verify the feasibility of the proposed method, which not only considers the control quality of system but also guarantees the service quality of the network.

An algorithm for OPEN proxcheme is proposed that functions in the two maximal decidable classes of first-order logic. The algorithm iteratively revises the formal theories according to new information and completes the evolution of the formal theories. The expressive power of the OPEN proxcheme is discussed. It is demonstrated that the OPEN proxcheme and R-calculus preserve the decidability of formulas in the two maximal decidable classes. Since the decision problem for modal logic and FO² can be transferred to that of the two maximal decidable classes, it can be concluded that the OPEN proxcheme and R-calculus are more expressive than the modal logic and FO². Practical examples, such as hardware description and verification, are provided.

To tackle the real-time scheduling problem of multi-load carriers in automobile assembly lines, a real-time scheduling approach was proposed based on Support Vector Machine (SVM). First, the real-time scheduling problem of multi-load carriers is formally described, and an objective function of assembly line output and the delivery distance is set up. Then, samples are generated by the simulation of the assembly line with material handling system to train an SVM model offline. Finally, the trained SVM model is used in the real-time scheduling process to make “wait” or “deliver” decisions. Experimental results indicate that the proposed approach outperforms the Minimum Batch Size (MBS) approach. It runs fast, performs well and to some degree adapts to the changes in the dynamic environment. It can be effectively used to improve the real-time scheduling of multi-load carries.

Software defect prediction is significant to the optimization of quality assurance activities. The Within Project Defect Prediction (WPDP) can produce high quality results, but requires historical data of the project, which is often not available in practical scenarios. The cross-project Defect Prediction (CPDP) can effectively overcome the drawback of WPDP. However, existing research suggested that CPDP is particularly challenging and often yields poor performance, and very few studies investigated the practical guidelines on how to select suitable training data for CPDP from multi-source project data. A novel multi-source data driven approach for CPDP is proposed. First, the hierarchical filter strategy based on the characteristics of both projects and modules is developed to select training data. Then, the Naive Bayes algorithm is used to realize the prediction model. Experimental results of 14 open-source projects show that the proposed approach significantly improves CPDP performance, and can compete with WPDP.

Based on vertex similarity in complex network and density peaks clustering, a community detection method is proposed. First, a vertex distance calculation based on vertex similarity and the shortest distance between vertexes is proffered. Then, the density peaks clustering method is applied to detect the community structure in network. The density peaks clustering method not only allows the detection of the community centers to establish the epicenter for community expansion, but also avoids the process of selecting parameters. The proposed method is compared with the classic algorithms on both real-world networks and synthetic networks. Experimental results demonstrate that the proposed community detection method is practicable and effective.

In order to achieve image fusion with high quality, a multi-focus image fusion scheme with co-sparse analysis is proposed. First, a co-sparse analysis operator learning model under regularization is developed, for which the analysis operator is obtained with similarity training data. Then, the images to be fused are represented by the salient features extracted with the analysis operator and their corresponding fusion coefficients are obtained with the choose-max rule. Finally, the fused image is reconstructed by minimizing a total variation norm regularized problem. Experimental results show that the proposed method outperforms the sparse synthesis based fusion method; the gain of the proposed method is up to 0.65 in terms of mutual information and 0.1 in terms of Q_ABF.

Based on the properties of connotation and denotation, classification boundary, classification granularity, and the relationship of class and individuals, the Thangka domain ontology is established by using the platform of Protégé. On the basis of this, the inference rules are imported and the information model ontology is built. The inference rules which are suitable to Thangka are designed based on Jena. The logical implication relationship between concepts and individuals of Thangka domain ontology is reasoned through inference rules. It the end, a case study of inference and query is carried out using a ritual apparatus which has special meaning in Thangka. The results verify the validity and rationality of the inference rules. An individual example of Sakyamuni is used to query the direction description and the indirect relationship, thus, the semantic query of Thangka knowledge base is realized.

For the requirements of interior fire location in high and large-span space buildings, an accurate 3D fire location method is proposed. To overcome the difficulty of establishing accurate mathematical model of camera in traditional non-linear calibration, an implicit camera calibration method based on Improved Particle Swarm Optimization (IPSO) algorithm and Least Square Support Vector Machine (LSSVM) is developed. First, the mapping relationship between the plane pixel coordinates and the spatial coordinates is established. Then, IPSO algorithm is applied to search the optimum parameters of the LSSVM regression model to improve the convergence speed and generalization ability. Finally the space positioning of the fire is realized by 3D reconstruction. The proposed method has been applied in fire location in high and large-span space buildings and the results suggest that the method is effective, fast and accurate.

The image gradient prediction method is applied to select and predict the trend of each pixel's gradient. The, the pixel's value is predicted according to the gradient and adjacent pixels. The use of the reversible data hiding algorithm based on prediction error can effectively reduce the prediction error and improve the hiding capacity. Moreover, in the hiding process, the embedding rule is adjusted to reduce the distortion to the host image. Experimental results show that the proposed algorithm can achieve a very embedding rate and has good concealment.

Based on the orthogonality between signal subspace and noise subspace, a new estimation algorithm of Direction-of-arrival (DOA) and range suitable for multi-frequency near-field source localization is proposed. By applying orthogonal transformation to the received data of transducers, a complex observation matrix is required. The eigenvalues and eigenvectors of the complex covariance matrix are computed using similarity transformation method of eigenvalue decomposition, and then the noise subspace is established. Considering the frequency spectrum of source signals, a suitable spectral function is constructed for multi-frequency signals to search for the locations of sources. The proposed method increases the applicability of localization system to multi-frequency sources. The actual localization system simulation results show the efficacy of the proposed method.

Most existing Direction-of-arrival (DOA) and polarization estimation methods, which are based on polarization sensitive array, generally assume that array element orient the referenced Cartesian coordinates strictly, that is no misorientation. However, in practical applications array elements have misorientations, and it is necessary to calibrate the misorientations. To solve this problem, a Taylor approximate expansion iterative method is proposed. It uses the first-order Taylor approximate expansion to calibrate the array elements repeatedly to get the array's misorientation matrix. The simulation results show that the estimation of DOA and polarization are quite close to the true values by using the proposed method.

In order to match different size block of Discrete Cosine Transform (DCT) algorithm, the architecture compatibility of different dimensional DCTs is implemented. A multidimensional DCT similar butterfly algorithm is proposed and its unit channel architectures are studied. First, based on the theory of DCT and tensor product operation, the theoretical principles of the DCT similar butterfly algorithm and its inverse transform IDCT are introduced, and the derivation of the multidimensional algorithm is given. Second, taking the arithmetical operation of DCT/IDCT stereo similar butterfly algorithm as the theoretical foundation, the one-dimensional stereo similar butterfly diagram is extended to multidimensional one. Finally, according to the relationship of one-dimension and multi-dimension, based on the one-dimensional channel unit architectures, the multidimensional unit channel architectures are put forward. Experimental results indicate that the proposeds algorithm only needs 50% adders and 30% multipliers of the traditional algorithm. The consuming time of this algorithm is related to the block size and dimension, for three-dimension and five-dimension, the consuming time of this algorithm is less than 10% of that of the common DCT algorithm. This algorithm is characterized by fast, low complexity and multidimensional compatibility.

A specific edge point detection algorithm for object localization is proposed. The algorithm can detect the edge information of interest and mark the edge points. The algorithm can get the location information of the target by analyzing the distribution of the edge points, and detect the specific gray scale pixel sequence and edge points of interest. For verification, the proposed algorithm is applied to detect the welding line and the unconstrained iris contour edge. Experimental results demonstrate that the proposed algorithm can accurately detect the edge points of the welding line. The algorithm is also applicable to locate the unconstrained iris contour. It obtains the edge points of the iris region first; then Hough Circle Transform is applied to vote a circle as the iris contour. The algorithm can locate the iris contours accurately for all kind of iris images in UBIRIS.v2 Database.

In order to solve the issue mismatching and structure disconnecting in exemplar-based image inpainting, an exemplar bidirectional matching algorithm for image inpainting is proposed. The neighborhoods of known and unknown regions in the repairing patch are investigated. The consistency criterion of the unknown region corresponding information in the matching patch and in the repairing patch is adopted. The priority of the repairing patch and the best matching patch are determined by the similarity of the known information and the consistency of the unknown information in the repairing patch. The known information of the image is sufficiently used. It can effectively solve the problems of mismatching and structure disconnecting in Criminis algorithm and the problem of blurring in Kawai algorithm. Experimental results demonstrate the effectiveness of the proposed algorithm.

The objective of this paper is to stabilize electronic image for aerial video with dynamic background. When shooting, as the camera follows the target, the video sequence changes scenarios. This is due to the movement of the camera and the change of background, so the scene is complicated. The difficulties in stabilizing video image with dynamic background include: an accurate estimate motion vector of camera movement and background compensation. To overcome such difficulties, this paper uses KLT track operator in combination with the M-estimator Sample and Consensus (MSAC) algorithm to realize feature point classification. Then, only the feature points on the background are used for global motion estimation to perform image stabilization. Experimental results demonstrate the feasibility of the proposed method.

This paper presents a Radial Time-frequency Peak Filtering (R-TFPF) method with adaptive Window-length (WL), that the structure tensor is used to adjust the WL of R-TFPF. The proposed method can be applied to identify seismic events and noise, therefore, we are able to process the noise with long WL to suppress the random noise more effectively and reduce the signal distortion with short WL. The performance of the proposed method is evaluated using synthetic data and field records. The results show that the method provides a superior performance in random noise attenuation and fine structure preservation of reflection events compared to the original R-TFPF method with fixed WL.

To overcome the deficiency of traditional quadrature detection algorithm in amplitude and phase extraction of controlled source electromagnetic detection with square-wave excitation source, an improved detection algorithm is proposed. The implementation process of the improved quadrature detection algorithm is analyzed, and the earth medium influence with high-pass properties is discussed. According to these characteristics, the matching low-pass filter and low sampling rate method are designed to ensure the detection precision of the improved algorithm. Experimental results indicate that the amplitude error of the improved algorithm is 0.01% ~ 0.72% compared with the sinusoidal-wave referenced quadrature detection, and the phase error is within 4°. The detection precision is improved at least one order of magnitude compared with the square-wave referenced quadrature detection algorithm. The improved detection algorithm is multiplication-free, in high real-time property, and is of low hardware requirement. Meanwhile, it can restrain the earth medium influence with high-pass properties. So the improved algorithm has good application prospect in controlled source electromagnetic detection.