With the analysis of preceding object motion feature and vehicle-borne radar measuring principle, a novel target motion model is established in the vehicle-borne radar coordinate system. This target motion model considers the relative motion of the intertial coordinate, the vehicle coordinate and the radar coordinate system. Also other specialness of ground vehicle were taken into account in the model, such as that the motion of ground vehicle is a 2D motion because of the limitation of ground surface, the vehicle mobility is small and the tracking coordinate system is moving. The whole motion states of the preceding target, including the longitudinal and lateral velocities, the longitudinal and lateral positions were estimated by the algorithm of adaptive Kalman filter, because it was difficult to determine the statistics of the system process noise and measure noise. Finally, road experiments, in which the host car was equipped with millimeter-wave radar and the preceding car was equipped with high precision automotive testing equipments, were carried out to verify the feasibility and performance of the estimation method. The results prove that the method can provide fine estimation accuracy, better filter convergence and stability.

As the conventional vehicle mechanical linkages are replaced by Steer-by-Wire (SBW) with electronic control, the driver is unable to sense the road information. A novel steering wheel torque and pinion angle feedback method is proposed with bilateral control structure. The design goal for SBW system is to allow the driver to match and set the dynamic relationship between steer angle and reactive torque input. Refer to electric power steering system, a steering force ratio is designed, which the road feel motor and steer motor control to improve the driver's steering feel according to the assistant control and return control methods that are used in the control logic. Vehicle experiments demonstrate that the proposed control methods can effectively feedback the road information to the driver, ensuring the vehicle's good returnability.

To predict, analyze and optimize the low frequency noise of a Chinese vehicle, a structural- acoustic coupling finite element model was established. The excitation signals at the suspension mountings and engine mountings, and the responding signal of the interior noise were measured with road testing. Appling the excitation signals to the structural-acoustic coupling model, the frequency response analysis was carried out and the low frequency noise was calculated. The simulation response of noise was compared with the result of testing. The consistency between the simulation and testing results proved the validity of the model. The driver's noise sensitivity on the thickness of each panel of the vehicle was calculated. According to the results of the sensitivity analysis, the damping layer was aligned to the structure to control the noise. The density and thickness of the damping layer were optimized by optimal experimental design. With the optimized damping layer, the peak value of the interior noise was reduced by 1.13 dB(A) and the overall sound level was reduced by 0.62 dB(A), the total mass of the damping layer was reduced by 1.953 kg.

According to the structural characteristics of wet clutches in hydraulic-mechanical transmission, the surface roughness, permeability of the friction material, and the centrifugal force of the lubricant oil film were investigated. Then a modified Reynolds equation based on the advanced average flow model was developed to calculate the relationship between oil film pressure and oil film thickness. The Green-Tripp method was used to develop the force and torque balance equations of single friction pair. The friction heat between friction plate and steel plate, and the cooling effect of oil film on the friction pair were analyzed. Thus, the change characteristics of the torque response, the film thickness and angular velocity during the engagement were acquired. Also the effects of the working parameters and material properties on the torque response during engagement were considered.

In this paper, first, the lightest structure under the strength condition was obtained via the implementation of strength optimization based on the wire-frame model of body structure and the sectional library. Sequencing optimization of the sectional library, local line search and local-global iterative process were contained in the strength optimization. Second, sensitivity analysis was carried out to select design variables, which are more sensitive to the stiffness and modal frequency of the body structure. Finally, the selected variables were optimized within the limits of sectional library to improve the performance of the body stiffness and modal frequency using the multi-objective genetic algorithm NSGA-II. This approach was implemented into the design process of an integral bus body. The results indicate that the proposed sectional optimization method is efficient and conducive to light-weighting of bus bodies.

In order to study the biomechanical response and injury mechanisms of the occupant pelvis during impact, a three-dimensional finite element model of the pelvis based the CT data of a 50th percentile Chinese male was developed. Then the experimental methods used by Guillemot et al. and Salzar et al. were adopted to validate the biofidelity of the model; the accuracy of the model was examined by simulation of cortical bone with hexahedral elements and simulation of trabecular bone with shell elements. Simulation results indicate that both mesh methods can be applied to simulate the pelvis impact responses and injuries, however, the results of the model with cortical bone meshed by hexahedral elements are more consistent with that of experiments. Moreover, the comparative analysis of simulation results with female pelvis under the same conditions reveals that geometrical differences between male and female pelvises affect their deformation resistant capacities in impact.

In this paper, the process of aerodynamic optimization for car body is divided into three steps. The first step is the two-dimensional body optimization aiming at global search; the second and third steps are three-dimensional body optimization without wheels and with wheels, respectively, aiming at local approximation and fine-tuning. Finally the corresponding evolution strategies of each step are developed according to the difference between the two- and three-dimensional cases. As the existence of wheels may change the aerodynamic behavior of the car body, so wheels in fixed position is taken as one of the constraint conditions, which may improve the effect of the optimization. Optimization results show that the drag coefficients of the vehicle with wheels are as low as 0.129 in cheese head style and 0.124 in socket head style. The dependability of this method is validated by the scaled model test in wind tunnel.

In this paper the vertical and lateral test models of vehicle body-frame double mass system were established. The inverse displacement analysis model of the excitation platform was put forward, which was based on the theory of homogeneous rotation matrix. The control command for the vibration exciter was solved via the inverse displacement analysis model. Meanwhile, a vehicle test for vertical natural frequency was conducted. The test result was consistent with theoretical value, which proved the accuracy of the test method.

According to the theory of ergodic property of stochastic process and the equivalence relation between time average of the energy consumption in velocity time series and statistical average of energy consumption in Velocity-and-Acceleration (VA) phase space, a model, which can calculate EV energy consumption based on VA probability distribution, was built up. Compared to the energy consumption model based on velocity time series, the efficiency of energy consumption calculation of the new model is greatly improved, meeting the demand of real-time calculation of EV energy consumption in urban route. The calculation accuracy of the model is validated by comparative tests between EV trial on the route of Changchun city and energy consumption simulation using the proposed model.

In order to enhance the credibility of the highway speed limit setting and ensure reasonable selection of speed control method, this paper analyzes the traffic flow data collected by single section and double section. The analysis is from the perspective of spatial and temporal characteristics of highway. Then three stability analysis models are constructed, which treat speed difference as the main research focus. Further, a recognition method of traffic flow stability is presented. The three new models are verified with the microwave monitoring data on a highway in Shandong province using WISSIM simulation software. The results show that identification model of traffic flow stability can identify the dispersion of traffic flow and describe the stability of traffic flow efficiently, which will provide reliable decision basis for highway management of dynamic speed limits.

A novel vehicle detection method based on Haar-like features and Adaboost algorithm is proposed to improve the capability of front vehicle detection of the driver assistance system. First, Haar-like features are selected from the training samples. Then, a learning algorithm based on Adaboost selects the efficient features from the Haar-like feature sets to yield vehicle detection classifier. The classifier is used to examine the testing samples by the pictures provided by the IEEE International Workshop on Performance Evaluation of Tracking and Surveillance. Result show that the proposed method can detect vehicles rapidly and effectively in daytime.

In order to describe staggered car following behavior in real traffic environment, a novel car following model was proposed based on Full Velocity Difference (FVD) model by incorporating the visual angle. The rate of change of the lateral separation angle is taken as a key parameter to describe the effect of lateral separation distance. The stability condition of the model was obtained using the linear stability theory. Then the characteristic of the model was investigated by theoretical analysis and numerical simulation. Results show that the lateral separation leads the stability of the proposed model smaller and the easy formation of stop and go traffic. The above findings have great practical significance to improve the stability of traffic flow and analyze car-following characteristics considering the lateral movement.

An improved public transportation assignment Logit model during peak period is proposed. First, the station arrest coefficient is defined and calculated. According to the transit stations with and without nonstop line and the crowding level, the stations are classified into three categories. The Logit model is improved from the aspects of the time function, sanction function and departure interval. Then, using a practical case, the influences of the parameters α and T_f in the model on the transit assignment are analyzed, and α and Tf are calibrated. Finally, error analysis of the transit assignment results is carried out, which validates the model.

In order to analyze the problem of selecting optimal path in traffic network, adaptive reliable shortest path problem is addressed in stochastic network. First, the mathematical model of stochastic traffic network is established to reflect the stochastic characteristic of traffic time in traffic network. Second, the optimal-reliable routing policy and optimal-reliable state chain based on reliability theory are uniformly defined in stochastic network; and the optimal-reliable state chain satisfies Bellman's principle that is the core of dynamic programming. Third, a successive approximation algorithm based on the dynamic programming is developed to solve the adaptive reliable shortest path problem in stochastic network, whose complexity is polynomial time. Finally, a computer program using Matlab language is developed to compute on the Sioux-Falls (SF) network. Numerical results in typical transportation network show the validity and feasibility of the successive approximation algorithm.

The physical and mechanical properties of asphalt mortars respectively mixed with 7% mine filler and with 7% diatomite were tested, the physical tests including point test, penetration test and Brookfield viscosity test, the mechanical tests including impact test and three-point bending test. The physical tests indicate that the diatomite can significantly strengthen the high and low temperature performance of asphalt. In addition, diatomite is more conducive to weakening temperature sensitivity of asphalt mortar. The mechanical tests indicate that diatomite asphalt mortar has better impact resistance performance and higher deformation resistance at low temperature. Thus, it is significant for preparation of diatomite modified asphalt concrete on the design of grade mixing ratio, so as to be conducive to promoting the use of diatomite modified asphalt mortar.

Elastic-plastic time-history analysis on a large span half-through Concrete-Filled Steel Tubular (CFST) arch bridge was conducted using ANSYS. Taking the acceleration record of El-Centro earthquake as the originally inputted wave, a series wave of different peak accelerations was produced. Then, the relationship between peak acceleration and following responses, such as the maximal displacement, strain energy of elements and ductility factor, were obtained. Finally, the dynamic stability ultimate bearing capacity and the aseismic behavior of the CFST arch bridge were determined. The analysis results have some theoretical significance to aseismic design of CFST bridges. The analysis shows that dynamic bucking of the structure is resulted from the development of both displacement and plasticity of the arch rib. The maximum displacement indubitably occurs near the arch crown and plasticity development always begins from the spring. It is concluded that the application with various responses is more necessary and reasonable in the research of aseismic behavior and dynamic stability of large span CFST arch bridges.

To explore new ways to optimize combustion under transient conditions, the influences of fuel injection parameters on the emission and combustion characteristics in a high-pressure common-rail diesel engine were investigated. The engine speed of 1650r/min was selected, and at this speed the engine torque was increased from 10% to 90% of the maximum torque within five seconds. Results show that, under transient condition, the smoke surges and NO_x is below that under steady-state condition, CA50 delays. When the injection timing is advanced, the smoke emission and NO_x emission increase, CA10, CA50 and CA90 are in advance. When the injection pressure is higher, the smoke emission decreases, NO_x emission increases, CA10 and CA50 are in advance. The feasibility of reducing distortion between transient-state and steady-state by adjusting the fuel injection parameters was also discussed. Results show that adjusting the injection pressure can reduce the performance difference between the transient-state and steady-state.

Based on the combustion chamber of a developed 2.0 L CR high-pressure DI diesel engine, a simulation model was established using software FIRE. The effects of different combustion chamber structures on the velocity field and turbulent characteristics inside the chamber were analyzed by using simulation. Besides, a dimensionless parameter Da number, which is closely related to the mixture formation and combustion process, was introduced in this paper. The influence of the chamber structure on the mixture formation mechanism, the combustion process and the generation rates of NO and soot were analyzed. The results show that the formation of NO and soot can be effectively prevented by controlling the mixture formation and the combustion process through the rational design of the combustion chamber structure.

The energy saving mechanism of the braking energy recovery system of Parallel Hydraulic Hybrid Vehicle (PHHV) is introduced. By the establishment of vehicle dynamics model, an AMESim model is set up correspondingly according to the actual parameters of the vehicle and its hydraulic components. Then, by simulation analysis of the vehicle braking and energy recovery, a corresponding curve is obtained. To verify the correctness of the simulation, experiments corresponding to different working conditions are carried out on a hydraulic experiment table. The experiment results are basically in agreement with the simulation results. By analyzing the causes of the error between the simulation and experiment, the following conclusions can be drawn. Under the conditions of shorter braking time and lower braking force, the energy recovery rate of the parallel hydraulic hybrid system is high, which is over 43.12%. The correctness of the simulation model is verified by experiment results. The AMESim model can be used to intuitively analyze the effect of braking energy recovery of PHHV.

The principles of hydraulic transmission were analyzed by illustrating off-grid wind turbine and on-grid wind turbine. A method to realize variable speed constant frequency function through the electro-hydraulic control by regulating the displacement of pump or motor was put forward. By modeling the off-grid wind turbine using Simulink/AMESim software, the method was validated through the cosimulation. Simulation results show that when wind speed fluctuates, the displacement of the pump can be adjusted to maintain the motor speed at the optimal tip speed ratio and the rotor maximum power capture can be ensured; meanwhile, the generator speed can be stabilized around rated value to output constant frequency electricity by adjusting the displacement of the motor.

A novel piezoelectric vibratory feeder was designed, which was driven by two vertically arranged rotary piezoelectric vibrators, where the inverse piezoelectric effect was utilized. By theoretical analysis, the nature frequency of the feeder system and the factors influencing the hopper's amplitude were obtained. The software Hypermesh was used to simulate the dynamics of the system, and the first three order vibration modes were achieved, from which we got the theoretical nature frequency of the system without damping to be 110.95Hz. Experiments on this feeder and on the same model cantilever piezoelectric feeder were carried out using M5 nuts as the conveying materials. From the experiments we got the relationship among the delivery speed, hoper amplitude, driven voltage and frequency. When the driven voltage was 240V and the frequency was 105Hz, the delivery speed of this new piezoelectric vibratory feeder was 0.23kg/min and the hopper amplitude was 110μm, respectively, which could meet industrial production requirement.

The volumetric error of a four-axis polishing platform with two aerostatic stages driven by linear motors and two rotational stages is modeled considering both position error and direction error by multi-body system theory. The individual geometric errors of the motion parts are measured by laser interferometer. The measuring results indicate that, for aerostatic stage, the position error shows neither linear increasing nor decreasing trend; the vertical straightness is larger than the position error and the vertical st4aightness, which is obviously different from the errors of traditional ball screw driven platform. The position component and direction component of the volumetric error of the resultant motion of two linear axes are investigated theoretically and experimentally considering the comprehensive influence of the individual geometric errors of the aerostatic stages. It is found that the vertical component of the volumetric error of the aerostatic stage is obvious. The causes of the geometric errors of the aerostatic stages are revealed, which provides the theoretical basis for geometric error compensation of the polishing platform.

A precision seeder for tray nursing seedlings of rice was developed using a new-type pneumatic vibration uniform-seeds device to promote the sowing performance of hybrid rice in low seeding rate. First, steady and uniform seeding condition with different working parameters of the uniform-seeds device was determined, including displacement-storage capacity ratio of air supply, A, setting angle, β, and trapezoidal half difference, C, of the sieving board. Then orthogonal tests were conducted to further investigate the effects of these parameters on the sowing performance, in which the seeding rates were 45 g per tray for supper hybrid rice, 65 g per tray for hybrid rice, and 85 g per tray for inbred rice, respectively. Results show that the best operational parameter combinations were: A was 0.036/40.0 C was 5 mm and β was -4°, 0° and +4° when seeding rates were 45 g per tray, 65 per tray and 85 per tray, respectively. Sowing performance tests based on the best parameter combinations were carried out. The sowing qualified rates were 88.33%, 90.83%, 93.06%, and the cavity rates were 1.76%, 0.56%, 0% for three seeding rates respectively. The sowing performance meets the requirements of precision seeding of super hybrid rice and hybrid rice in low seeding rates.

The foot-ground contact biomechanics of German Shepherd Dog (GSD) in normal walking, trotting and jumping gaits were investigated using a pressure plate system. The peak of the vertical ground reaction force and the dynamic foot pressure distribution on the paw pad of right-forefoot was measured and analyzed. It was found that the peak ground reaction force acting on the forefoot was about 0.69 times of the weight of GSD in walking gait; while in jumping gait, the reaction force increased to 1.43 times of weight of GSD. Meantime, along with the increase in the vertical foot-ground impact force, GSD made adjustment of its foot contact region for force bearing. In walking gait, the contact region was the 4th and 5th toes, and in jumping gait the contact region was shifted to the 3rd, 4th toes and metacarpal. Such adjustment is of great benefit to cushion and impact energy absorption.

Numerical simulation is carried out to study the drag reduction characteristics of porous bionic jet surface according to RNG k-ε turbulence model. The results show that there exists a linear dependency of the drag reduction on the rate of velocity, that is, the higher the rate of the velocity, the better the effect of drag reduction. The maximal efficiency of drag reduction is 59.02%. Friction coefficient decreases at first, and then increases along with the central line of the bionic jet surface with single hole. The efficiency of drag reduction in local region can be up to 111.8%. Furthermore, the more the number of jet holes, the better the drag reduction effect. The mechanism of the drag reduction of the bionic jet surface is studied. It is found that the jet-flow increases the thickness of the viscous sublayer in the boundary layer by changing the structure of the flow field in the vicinity of the jet surface. As a result, the gradient of normal velocity, perpendicular to the surface, is decreased, which contributes to remarkable drag reduction. Meanwhile, the gradient of normal velocity generates streamwise vortex structure and develops small secondary vortex on the surface, which results in inhibiting momentum exchange between fluids, thus, reducing the friction on the jet surface.

There exist several problems in the production and stress analysis of bogie frame of railway passenger cars. It is very difficult to establish the dynamic welding heat source model and implement the processing. To solve the problems, the theory of space welding path is investigated and the shifting mode of the welding heat source coordinate is defined. Then, the welding heat source coordinate model is established by parameter inputting and operating in software ANSYS. A loading method of the plane curve path, space curve path and bogie frame of passenger car model is proposed. Numerical simulation is used to perform the moving heat source model and welding curve path method, thus to understand the welding process and control the residual stress deformation by choosing reasonable production process. The welding process simulation can be realized efficiently and exactly. Therefore, the theory proposed is of significance in theoretical research and engineering application.

The error of traditional fatigue life algorithms of materials is big. To overcome this problem, a model of the relationship among stress concentration factor (K_t), stress average (S_m), stress amplitude (S_a) and median life (N50) was built using BP neural network combined with genetic algorithm. This model is based on the classification of fatigue life data of materials, and it can be used to predict fatigue lift of materials with finite life data. Experiment results show that, using this model the relative prediction error is below 5%, in another word, the prediction accuracy of this model is higher than that of the traditional prediction algorithms. So this model can be applied to prediction the fatigue life of materials.

This paper investigates the effect of partitioned cooling on the mechanical strength and hardness of hot forming steel based on the flat tool cooling test. An exponential model of the mechanical strength-hardness-cooling rate was established using the dimensional analysis method, and this model was validated to be effective. By writing the user subroutine, this model was introduced into the prediction program in the self-developed software KMAS_HF platform to validate the hardness and strength prediction of U-shape hot forming product. Results show that the model is reliable and applicable in predicting the mechanical properties of hot forming products, which can be used to guide the optimization of hot forming process and the design of hot forming tools.

A flexible 3D stretch-bending process was developed based on the idea of multi-point forming technique. To realize precise 3D shape of the profile, the freedom of single multi-point die in the process of 3D stretch-bending was analyzed first, and the structure of fundamental unit with 4-DOF was designed. Then, the calculation method of unit adjusting parameters was presented, and a combination of serial and parallel adjustment control method was proposed. In addition, the mathematical model of the unit was set up. With the structural component of high-speed trains as an example, 3D stretch-bending test was carried out. The experimental results show that the flexible 3D stretch-bending equipment meets the practical requirements. Compared with manual adjustment, automatic unit adjustment can improve the production efficiency by 60%, as well as reduce labor intensity.

To reduce the influence of module joint actuator and sensor failure on the controllability of the reconfigurable manipulator, an online fault estimation method based on the decentralized fault-tolerant control scheme was proposed. Based on the modularized property of the manipulator and Lyapunov stability theory, a decentralized adaptive sliding mode observer is constructed to estimate the actuator and sensor faults adaptively. Meanwhile, the nonlinear terms are approximated or compensated by neural networks. Then, the idea of non-singular fast terminal mode control was employed to achieve decentralized fault-tolerant control. Finally, numerical simulation was carried out for two 3-DOF reconfigurable manipulators with different configurations. The results demonstrate the effectiveness of the proposed fault-tolerant control scheme.

In order to improve the robustness of load disturbance and parameter uncertainties of the Permanent Magnet Synchronous Motor (PMSM) speed control system, first, a nonlinear integral sliding surface was designed based on the fractural order calculus of the speed error. Then, a robust Fractional Order Integral Sliding Mode Control (FOISMC) algorithm was proposed for the PMSM speed control system. The stability of the proposed FOISMC method was proved by the Lyapunov stability theory, and the performance of the proposed controller was analyzed. Analysis and numerical simulation results show that the proposed method has better dynamic performance, robustness and speed tracking accuracy than the conventional integer order Integral Sliding Mode Control (ISMC) method and PI control method.

To meet the special requirement of switching power supply used in coal mines, a switching power supply with a wide range of input voltage is proposed based on a three-level converter. First, a small signal model of the converter is derived using the state space averaging method according to the operating principle of the converter, and the transfer functions are obtained from the small signal model. Then, with the transfer functions, a closed-loop control method is established and a passive lead compensation controller is designed. Finally, a simulation model of the three-level converter is developed to test the proposed converter. Simulation results show that with the input voltage between 95V and 825V and varying load, the output voltage can be controlled to be stable at a specific value.

Based on the analysis of the Autonomous Aerial Refueling (AAR) technology, a Flash LIDAR-based AAR scheme was proposed. This scheme can reduce or eliminate the dependence on GPS for AAR. First, according to the flight formation characteristics of the tanker and receiver aircraft, the horizontal and vertical tail of the tanker was selected as the characteristic plane, and a feature extraction algorithm for the tanker's point cloud subset was worked out. Second, Lagrange multiplier method was employed to calculate the unit normal vector of the best fitting characteristic-plane. The unit normal vectors of the successive point cloud frames were available data for visual odometry to estimate the relative pose change between the tanker and receiver aircraft. Simulations show that the feature extraction and pose estimation algorithm is feasible. Finally, in order to improve the accuracy of relative pose estimation of visual odometry, vector median filtering, horizontal and vertical tail correlation algorithm were employed to eliminate measurement errors. Results show that the filter algorithm is feasible in the condition of white noise SNR 25 dB.

We introduce the model-based diagnosis into local pinpointing of ontology debugging. The concept of MUCS is given on the basis of component terminology definition to achieve the local debugging on terminology. We take the advantage of MUCS to pinpoint the logical errors of unsatisfiable concept to the axiom interior, which is independent of the existing ontology debugging. One MUCS for an unsatisfiable concept is calculated by calling an external DL reasoner, and all MUCSs are obtained by combing the principal of breath-first with hitting set theory of Reiter. Finally, the realistic ontology is used to conduct a comprehensive test of the MUCS algorithm. Experimental results show that the MUCS debugging algorithm could point out the local position of logical conflict in an axiom specifically, and efficiently and accurately complete the ontology debugging task.

Based on users' historical trajectory data, users' regular routes were mined and extracted. A concept of support route was defined after route pre-processing and grouping. A regular route mining algorithm was proposed based on the support score. A feature of Regular Stop Rate (RSR) was used to improve the accuracy of the transportation mode recognition. The effectiveness of the approach was validated based on the GPS data of 178 users over four years. A real user study was also performed among 37 users. The experiment results demonstrate that the algorithm can effectively extract the regular routes and is robust to slight disturbance in trajectory data.

In this paper, first, the influencing factors of Wireless Sensor Network (WSN) lifetime are summarized. Then, a 4-layer sandwich model for service-oriented WSN lifetime assessment is established. Finally, a generalized lifetime assessment method was proposed. The novelty brought by the sandwich model is that it defines a general lifetime decision criterion, which is suitable for the diverse WSN applications; and it can be used to incorporate their scenario specific requirements into quantitative criteria of lifetime assessment. Compared with existing lifetime assessment model, the generalized assessment method based on 4-layer sandwich model is more flexible, reliable and service-oriented.

In order to support the sleep-wakeup functionality of MOST50 network, a sleep-wakeup scheme for MOST networks is proposed. This scheme includes the hardware designing works for network nodes supporting sleep-wakeup, the network wakeup scheme design was implemented through the time-sequence of Electrical Control Line (ECL) pulses, the implementation of network sleep through the queries and acknowledgements of massages between nodes. The validity of the proposed sleep-wakeup scheme for MOST networks is demonstrated in the application of MOST50 network.

Using computer vision techniques to achieve early fire detection is a hot research topic. In this paper, a candidate flame pixel classification algorithm is proposed based on color model. First, the features of the flame region in RGB and YCbCr color models are introduced. Then YCbCr model is used to build flame pixel detection model. In addition by converting the rules based on RGB color model into corresponding rules in YCbCr color model, new rules in YCbCr color model are developed, which can further alleviate the harmful effects of changing illumination. The method can be used for real-time fire detection.

In order to solve the problem of high computational complexity and low matching speed of the traditional Scale Invariant Feature Transform (SIFT) algorithm, a new method of feature description using Local Binary Patterns (LBPs) was proposed. This method was used to describe the zircon image feature points detected by SIFT algorithm. The dimensionality of the feature vector was reduced using Principal Component Analysis (PCA), and the descriptor matching was carried out with Euclid Distance. The new description method of LBPs has a more simple calculation, and this algorithm has excellent features of rotation invariance and illumination invariance. The descriptor matching is faster after PCA dimension reduction. Experimental results show that this method meets the requirements of automated sample search, and significantly improves the matching speed of zircon images, improves the efficiency of the instrument.

A new index structure, DSD⁺, which is the dynamic sketch with direction index, is proposed to solve the problem of aggregate query of moving objects in the road network environment. This technique combines the sketch technology, which can solve distinct counting problem, and the AMH+ technology, which can give out efficient approximate aggregate query results with smaller storage space; and utilizes the characteristics of aCN-RB-tree that supports aggregate by the direction property. The performance analysis and the experimental results show that, with the same query time and precision, the query storage space of the proposed DSD⁺ index structure is less than that of existing index structures.

A new fingerprint matching algorithm is proposed in this study. First, the k-nearest neighbor and ridge line tracking methods are used to construct new topology structures for fork point and end point, respectively, in order to extract detailed characteristics of ridge line, neighboring region and minutiae. Then, a fast secondary matching process is implemented using the techniques of feature weighting, coordinate transformation, and semi-variable bounding box. Experiment results show that the proposed algorithm can run fast and has low false reject rate and false accept rate.

An online sorting method of irregular potatoes is proposed. First, the R-component image is extract, and then the binary image and boundary image are obtained by thresholding method. Second, thirteen essential geometrical features, such as eccentricity, rectangle degree and roundness, and ten Fourier descriptors are extracted. Third, the geometrical features of the potato image sample are feed into the module of I-RELIEF algorithm, which exports a weight for each feature. Fourth, the features of the potato image sample with weights are feed into the training module of Support Vector Machine (SVM). Finally, the classifier model is used to make decision and achieve the grading result online based on the potato's features and weights. Results show that the SVM method can detect and sort four potatoes per second with the help of I-RELIEF module, and the overall accuracy is 98.1%.

Under the condition that transmission power and relay power of the nodes are unknown, an opportunistic relay selection algorithm is proposed to solve the problem of best relay selection in AF-MIMO cooperative communication system without direction link. By theoretical analysis and mathematical derivation, the effects of the number of cooperative relays and the power allocation in each part of the system on the channel capacity are investigated. A simple criterion is derived using the instantaneous channel state information and the best relay is selected in terms of the maximin criterion, which improves the overall channel capacity without increasing the computation complexity. Simulation results show that, the relays selected using the proposed algorithm can significantly improve the channel capacity, ahich confirms the effectiveness of the proposed algorithm.

The errors of the basis function in the receiver, caused by the difference of the spectrum environment between the transmitter and receiver of Transform Domain Communication System (TDCS), will degrade the system performance significantly. A design method of the basis function, which has suboptimal solution of the Signal-to-Noise Ratio (SNR), is proposed to solve this problem. Based on a detailed study of the data sending and receiving process, the optimized Bitter Error Rate (BER) of the system and its relationship with the basis function of the receiver is discussed. In the case that the system with spectral mismatches can not obtain the optimal basis function, a design method of the basis function similar to equal-gain combing is proposed, and then this basis function is used for modulation. Simulation results show that, in the conditions of both multi-user and single-user with high spectral mismatch ratio, the receiver is suboptimal when the basis function is designed similar to equal-gain combing. In the single-user condition, the BER performance improvement of suboptimal reception is approximately 4 dB compared with the reception method when the mismatch ratio is 40%, and the SNR loss is below 1.5 dB at the spectral efficiency of 80% compared with the optimal reception.

A new filtering algorithm is proposed. In this algorithm, the probability density function is generated by the sigma point H_∞ filter. The sigma point H_∞ filter has very high accuracy and strong robustness to uncertain observation noise, and the filter algorithm integrates the new observation. So the generated probability density function can reasonably approximate the real posterior probability distribution of the system state. At the resampling step, the degeneration problem is effectively overcome and the accuracy of state estimation is improved by using the quasi-Monte Carlo-based resampling algorithm. Simulation results demonstrated the feasibility and effectiveness of the proposed algorithm.

To improve the unreliable Doppler frequency estimation under the condition of extremely low SNR and high dynamic state, a searching range prediction method was proposed based on Doppler frequency shift estimation used in spread spectrum systems. Using Doppler frequency variation limitation over adjacent two received signals, the proposed method narrows the searching range of the later estimations from the previous estimation results, thus, enhancing the estimation reliability. Simulation results show that this proposed method can effectively increase the Doppler detection reliability.

A new method for named entity recognition in Chinese medical records based on cascaded Conditional Random Fields (CRFs) is proposed. The first layer of the cascaded CRFs is used to identify the basic named entities of body parts and diseases. Then, the identified results are fed to the second layer for recognition of nested named entities for complex diseases and clinical symptoms. A new combination feature, composed of part-of-speech features and named entity features, is defined. This new feature together with the character features, word boundary features and context features in a sentence are taken as the feature set of the second layer. In the experiments based on CRF++, the proposed method yields a 3% higher F-score than cascaded CRF without the combination feature. Moreover, compared to single layer CRF method, it yields a 7% higher F-score, a significant increase in overall performance.

To enhance the accuracy of multi-source image fusion, an adaptive fusion method is proposed. This method is based on the area objective assessment in Finite Discrete Shearlet Transform (FDST) domain. The source images are decomposed to subband images with multi-scale and multi-direction by FDST. The gradient information correlation factor is taken as the coefficient weight for low-frequency subband fusion; while for high-frequency subband, consistency selection of the coefficient absolute value and area standard deviation is adopted as the fusion rule. The fused low and high frequency coefficients are reconstructed to image by Finite Discrete Shearlet Inverse Transform (FDSIT). Fusion experiment is done with several sets of different modality images, and objective performance assessments of the fusion results are implemented. Results indicate that the proposed method performs better in subjective and objective assessments than other existing Multi-Scale Decomposition (MSD) fusion techniques.

In this paper, first, the fast computation algorithm of Zernike moments proposed by Mukundan and Ramakrishnan was amended. Second, the transformation method from pseudo polar coordinates to the Descartes coordinates was proposed. Third, the effects of pretreatment of image on both original algorithm and the amended algorithm were tested. For each image at fixed order and fixed repetition, four moment magnitudes corresponding to the four rotations were obtained. Then the ratio of standard deviation and mean value of the four magnitudes were taken as the evaluation indicator. For comparison, the original algorithm and the amended algorithm each got 5×11 indicators from five selected images. Results show that the amended algorithm has 44 indicators exceeding that of the original algorithm.

The RNA interference (RNAi) technology is used to regulate the expression of cycloartenol synthase in P. ginseng hairy root. The ginsenoside was regulated successfully at the molecular level. The heritable ginseng hairy roots, whose total ginsenoside content was increased by 82.17%, were obtained. Meanwhile, the study results show that RNAi can be successfully applied to production regulation of plant secondary metabolites.