When a vehicle passes over a bump or a rough road, the impact torque on the steering wheel will be produced. Over large impact torque will bring unpleasant to the driver. The impact torque on the pinion gear is estimated based on the established dynamic model of Electric Power Steering (EPS). Then, the compensation current of disturbance from the road wheels is determined by a control algorithm to reduce the impact torque on the steering wheel. The vehicle test results show that the control algorithm can effectively reduce the impact torque on the steering wheel without increasing elements of the EPS system and it has potential application value.

In order to improve the effect of multi-step forecasting for short term traffic data collected from the loop in Sydney Coordinated Adaptive Traffic System (SCATS), on the basis of data preprocessing, a bi-level model of multi-step forecasting using dynamic neural networks was designed. This model includes a multi-step forecasting method based on Nonlinear Autoregressive model with exogenous inputs (NARX) neural network and a predictable steps online estimation method based on Focused Time-Delay (FTD) neural network. Validation and comparative analysis were carried out using data of loop in SCATS measured from a megacity. The results indicate that the proposed method can further reduce the errors of short-term traffic multi-step forecasting in SCATS.

On the basis of existing traffic status discrimination under normal conditions, considering the influence of snow-ice disaster on road capacity, real-time discrimination methods of section and intersection under snow-ice weather was proposed. Through considering factor such as road level, section capacity and section length, determined traffic status weights of sections and intersections, achieved the real-time discrimination of regional traffic status. The method was verified using TransModeler microscopic simulation software. The results showed that under snow-ice weather the method has good performance in aspects of accuracy and prediction time, it provides new ideas for traffic state discrimination field under snow-ice conditions.

Smart cameras in detection of mixed traffic flow parameters are in shortage. This creates a huge load of communication between the cameras and monitoring center. To solve this problem, this paper proposes a method of hardware and software codesign and implementation of a smart camera platform to detect mixed traffic flow parameters. Charge-coupled Device (CCD) camera sensor and Digital Signal Processor (DSP) are used to construct the camera hardware prototype. The video processing methods including background initialization, background model, foreground obtaining, image segmentation, pattern extraction, multi-classification etc. are used to construct the software system. Experiments were conducted under different mixed traffic conditions to test the camera system. Results show that this system can accurately obtain mixed traffic flow parameters. This study provides a reference for the development of smart camera system for mixed traffic detection.

To reasonably match machine power under excellent compaction quality, parallel hybrid power system for tandem vibratory rollers and its control strategies were put forward based on the analysis of cycle load characteristics of the rollers. Taking the hybrid driving system of a tandem roller as an example, the load cycle from high speed starting to braking and stopping was investigated using AMESim and ADAM simulation. Simulation results show that the engine speed fluctuation is cut down more than 50% in starting and stopping process, the driving speed is stable in compaction process. The engine output torque is limited to 150 Nm and the braking energy is effectively recovered, which keeps the balance of the charge of power cells in a working cycle. It is demonstrated that the parallel hybrid power system of tandem vibratory rollers and its control strategies are correct, the power is matched more reasonably with higher utility rate and compaction quality compared to original hydraulic system.

Based on the measured temperature, meteorological and strain data of the auxiliary shipping channel bridge of Sutong bridge during two years of operation, the effective temperature and strains of concrete box girder were studied. The results show that the smaller size of the concrete box girder, the larger range of the effective temperature. For the bridge with design reference period of 100 years, the ranges of effective temperatures of the concrete box girders at the pier top and at the middle span are (-4.3 ℃, 37.3 ℃) and (-6.1 ℃, 42.2 ℃) respectively. The correlation coefficient between effective temperature and average high (or low) air temperature for 3 consecutive days reaches 0.97, and the equation obtained by regression analysis is quite useful in estimating the extreme effective temperature of the concrete box girder. At last, a method to modify the effect of concrete shrinkage and creep was proposed, and the relative strains at the bottom section of the main pier and at the bottom slab of the box girder adjacent to the main pier were predicted using the effective temperature of the concrete box girder.

By bending test of asphalt mixture under freeze-thaw conditions,the influences of freeze-thaw cycles,asphalt-aggregate ratio and gradation composition on the flexural characteristics of asphalt mixture were investigated.Experiment results show that,the flexural tensile strength, the flexural-tensile strain and the stiffness modulus of asphalt mixture decrease with the increase in the number of freeze-thaw cycles. Obvious influences of the asphalt-aggregate ratio and gradation composition on the flexural characteristics were observed.As the asphalt- aggregate ratio increases, the influence of freeze-thaw cycles on the flexural characteristics decreases.The flexural characteristics of AC-13 mixture are better than that of other two gradations.It is shown that applying AC-13 mixture with 5.5% OAC or higher can improve low temperature flexural characteristics of the asphalt mixture under freeze-thaw cycle condition.

The projective of this research is to investigate the effect of horizontal aggregate distribution in asphalt mixtures on splitting test. Based on the aggregate homogeneity index and through loading in different directions of the horizontal cross-section of asphalt mixtures, the effect of horizontal aggregate distribution on the maximum stresses (in-plane principal stress and tensile stress) of the splitting test was numerically simulated using Finite Element Method (FEM) and verified by real test. Results show that the effect of horizontal aggregate distribution on the maximum stresses of the splitting test is insignificant and there is no obvious correlation between them. However, the correlation between the horizontal aggregate distribution and the variation of the maximum stresses is significant.

The effects of dynamic fluctuations of needle movement and hydraulic pressure on flow characteristics inside the nozzle hole and atomization characteristics outside the nozzle hole for multi-hole injector were investigated by numerical simulation. Results demonstrate that the flow characteristics and atomization characteristics change significantly with respect to the actual fluctuations of hydraulic pressure and needle movement inside the nozzle. The spray morphology and spray characteristic parameters, such as penetration distance and spray angle, are more consistent with the experimental results with the introduction of dynamic fluctuations of hydraulic pressure and needle movement inside the nozzle. The penetration distance is obviously larger and the spray angle is relatively smaller at the early injection stage when the hydraulic pressure and needle movement are taken as constants. Therefore, the influences of dynamic fluctuations of needle movement and hydraulic pressure on fuel flow inside the nozzle, atomization outside the nozzle, as well as on mixing, combustion and emission generation should be fully taken into consideration in diesel engine simulation.

Droplet swarm of diesel-gasoline blend fuel is prepared on an experimental system based on ultrasonic atomization technology. The mean droplet diameter, air/fuel ratio and droplet speed are independently adjustable to simulate the flow and mixing conditions of the spray in the cylinders of internal combustion engine. The droplet swarm is imported into a Controllable Active Thermo-Atmosphere Combustor (CATAC) to investigate the steady combustion characteristics of the droplet swarm. Results show that changing the diesel-to-gasoline mixing ratio, the fuel feeding speed and air flow leads the shape of spontaneous flame changes significantly. As the diesel component increases in the blend fuel, the lift-off height of flame increases, while the width of the flame decreases. The variations of fuel feed speed and air flow influence the equivalence fuel-to-air ratio and the central jet velocity, thus greatly influencing the fuel combustion characteristics.

A fuel-air coordination control strategy is introduced and applied to control the transient performance under Low-Temperature-Combustion/Compression-Ignition (LTC/CI) combustion mode switch. The experimental results indicate that, with this strategy, stable operation condition is established quickly and smoothly after combustion mode switch, thus achieving low emission level benefit of LTC.

To improve the system efficiency and fuel economy of hybrid power loader, parameter matching method for the parallel hybrid power loader was proposed. First, the system structure, operating conditions and control strategy of the hybrid power loader were analyzed. Then the objective function and constraint function for optimization were established. Parameter matching was implemented on a 5 ton hybrid power loader. Particle swarm optimization algorithm was employed to optimize the transmission system. Simulation results show that, after parameter matching, the system efficiency was improved and fuel consumption was reduced, which could be further reduced by system optimization.

To increase the mean time between failures of the over-running clutch of dual-turbine torque convertor of a loader, a new type of over-running clutch was proposed based on the analysis of disadvantages of the original structure. The stresses and strains of the original structure and the improved structure were calculated using Finite Element Method (FEM). Results show that, for the improved structure, the lock angle induced by the abrasion of the inner ring raceway changes little, and under separation state, so it improves the working reliability of loader.

The welding strength of a stamping welded hydrodynamic torque converter is calculated based on the Fluid-Solid Interaction (FSI) analysis technology. The calculation is based on the structural feature of a welding seam in stamping welded hydrodynamic torque converter. It applies contact algorithm, implements an accurate prediction of the welding strength of turbine by combining CDF and FEA. The distortion and stress distributions of the turbine structure are obtained. The results show that, under typical braking condition, the welding strength of stamping welded hydrodynamic torque converter can meet the requirements. This work provides an important reference to the structural design and strength analysis of stamping welded hydrodynamic torque converter.

The stress distributions of three types of joints, namely balanced single-lap bonded/ bolted hybrid joint and corresponding bonded only joint and bolted only joint, were investigated and compared by gradually developing accurate three dimensional Finite Element Analysis (FEA) models. First, the FEA model of bonded only joints was built and employed to obtain the adhesive stress of the joint, which was in good agreement with that obtained by theoretical model in the literature. Then, the bonded only joint FEA model was extended to the bonded/bolted hybrid joint model. The adhesive stresses in the hybrid joints with two kinds of adhesives were compared with that in the corresponding bonded only joints. The load sharing ratios by the adhesive layer and by the bolt of the hybrid joints were calculated. Results show that the adhesive stress distribution of the hybrid joints with phenolic resin adhesive was almost the same as that of the corresponding bonded only joints; the bolt shared very little load. However, in the hybrid joints with acrylic resin adhesive, the adhesive stress decreased significantly and the bolt shared large load. Finally the FEA model of the bolted only joint corresponding to the hybrid joint was built. Comparison shows that hole stress concentration in the hybrid joint is much lower than that in the bolted only joint, which means the hybrid joint can effectively eases stress concentration. In hybrid joints with low modulus adhesive, load can be shared by the bolt and the adhesive. So the performance of the hybrid joints is better than that of traditional joints.

We present a new time-varying mesh stiffness model using polynomial function with regard to the time-varying mesh stiffness of the high-bevel gear of a radar device. By analyzing the revolution speed, load torque and damping factor, we build the mesh stiffness model and the nonlinear dynamic equations of the high-bevel gear. We compare the proposed mesh stiffness model with traditional rime-varying mesh stiffness model of multi-order harmonic superposition and reference data model of time-varying comprehensive stiffness. It is shown that, with small damping and overload, the proposed model is much closer to the actual characteristics of the time-varying stiffness. This model lays a foundation to study the gear transmission dynamics, such as vibration damping, noise reduction, smooth transmission and other aspects with small damping and overload.

Due to the huge size and high temperature, it is a difficult task to on-line measure the geometrical parameters of large-size forged components. To solve this problem, a dimensional measurement system for large-size forged components was developed based on planar Charge-coupled Device (CCD), which can be used to for remote and contactless measurements of various geometrical parameters of high temperature forged components. The focus of this article is placed on coaxiality measurement of multi-diameter forged shaft. The mathematical model was established using planar CCD imaging model. Experiment results show that the measurement error is less than 1.5 mm for general size gorging (400-550 mm in diameter), which means the relative measurement error is about 0.3%. Compared with existing measurement systems, the proposed system has many advantages, such as simple operation, convenience and high reliability etc. It is estimated that about 7% of raw materials can be saved using the proposed system.

The inertia dynamometer is an important test equipment for the main shaft brake of wind turbine generator system. Currently, the capacity of large-scale wind turbine generator system has already reached 6 MW in China, and existing inertia dynamometers can only complete the 3 MW or lower test. This article introduces the latest development of the inertia dynamometer for main shaft brake of 6 MW large-scale wind turbine generator systems. The working principle, structure and solutions for key technology of the inertia dynamometer are described. This inertia dynamometer system is not only be used for wind turbine brake, it is also used for industrial brake, heavy truck brake and marine windlass brake testing, thus increasing the investment benefit. Primary test results show that this inertia dynamometer fulfils functional requirements.

A multibody dynamic model of roller chain system was developed in SimMechanics by considering that components are interconnected by flexible joints, and sprockets have complete tooth profile. Based on this model, the generalized forces of the flexible joints and the dynamic equations of the system were obtained using the principle of virtual work and Lagrange methodology. Considering complete tooth profile, the distributed element contact model was established. By introducing contact-pairs index matrix and local contact coordinate systems, a fast method for contact search, which is based on homogeneous coordinate transformation, was presented. The contact-impact forces between roller and each element of tooth profile were achieved by simulation. The proposed method is validated by results comparison with ADAMS simulation.

Polycrystalline pure copper was irradiated by High Current Pulsed Electron Beam (HCPEB) to investigate the effect of irradiation on pure copper. Both the surface microstructures and corrosion properties of pure copper before and after HCPEB treatment were studied in detail. Experimental results indicate that the corrosion resistance of pure copper irradiated by 10 pulses was remarkably improved according to the analysis of polarization curves of copper in seawater and electrode impedance spectroscope. Transmission Electron Microscopy (TEM) observations show that large amount of supersaturated vacancy defects were produced when the material surface was subjected to HCPEB irradiation. The agglomeration of vacancies can cause the formation of vacancy clusters, such as vacancy dislocation loops and Stacking Fault Tetrahedrons (SFTs). The vacancy clusters on the irradiated surface promotes the formation of a thick and dense passive surface layer, whose ability to bock the corrosive anions is enhanced, thus improving the corrosion resistance of the material.

When aquatic creatures move in fluid medium, the elastic deformations of their epidermis vary with the fluid load. The deformations couple with the special non-smooth structures of the subcutaneous tissues under certain conditions to form special coupling function surfaces. Based on the above biological coupling phenomenon and bionics similar principle, a new design concept of dual factors coupling functional surface caused by form and flexible material is proposed. This kind of surface is called Bionic Coupling Functional Surface (BCFS). The BCFS is composed of two layers of materials with different properties. The deformation of the surface layer and the coupling function between surface layer and the subcutaneous layer with non-smooth structure are utilized to actively control the fluid medium and achieve efficiency enhancement. The implementation of the BCFS on typical fluid pump is explored. Experiment results show that with BCFS the pump efficiency is increased by 5%.

As an auxiliary tool for traveling, hiking stick is very useful for travelers, in particularly for desert hiking. A bionic desert hiking stick was developed inspired by the structure, morphology and excellent travelling mechanism of ostrich foot on sand. This design has a stick tray on the tip that mimics the foot pad of ostrich, which possesses excellent characteristics of san fixation and anti-sinkage. The structure of the bionic stick tray and the contacting area between the tray and sand could be automatically adjusted with the changes of vertical load of the stick and the rigidity of sand. Using Discrete Element Method (DEM) software PFC2D^®, the interaction between the bionic stick and sand at three different work states, including the initial contacting sand state, the middle load bearing state and the ultimate load bearing state, were numerically investigated. Simulation results show that, compared with common desert hiking sticks, the bionic hiking stick possesses the excellent characteristics of sand flow fixation, anti-sinkage and high load bearing capacity.

To reduce the frictional resistance between the barrel and the rubber seal ring of shell body during the launching process of aerodynamic extinguishing cannon, a bionic drag reduction technology of dimpled surface was applied to the rubber ring. The motion model was established for the dimpled drag reduction surface of the rubber seal ring. The effects of the characteristic diameter of bionic dimples on the rubber seal ring and the shell speed on the frictional resistance between barrel and shell body were investigated by numerical simulation. Results show that, at the same shell speed, with dimple characteristic diameter of 2 mm, the highest drag reduction effect, 17.191%, was obtained; with dimple characteristic diameter of 4 mm, the lowest drag reduction effect, 3.158, was obtained. It is revealed that the mechanism of the drag reduction is that, during moving inside the barrel, the lubricant stored in the dimples on the seal ring of shell spills over to lubricate the barrel wall and seal ring, which reduces the frictional resistance and thus increases the launching speed.

A control system of a mini-weeding-robot platform was developed based on C8051F340 single-chip microcomputer and tested in paddy field. The design process of the hardware and the flow of the software were described. The control system was composed of three module, the onboard control system module, the data translation and transmission module on PC and the remote control module. The remote control module can complete automatic navigation control with the image information and manual remote control. Field test results show that the mini-weeding-robot platform can walk in the paddy field with the speed of 0.2 m/s, eight-degree-of-freedom movement, and rapid response speed of 30 ms. The clean rate was 81.3% and the injury rate of rice seedlings was 3.6% using mechanical touching weeder.

After considering the factors of production, ecology, economics, meteorology, society and soil, an index system for selecting water-saving irrigation technology was established. An optimization model of water-saving irrigation technology was developed based on Ant Colony Algorithm (ACA) and Support Vector Machine (SVM). The ACA was used to select indicators and SVM was used to build the classifier. Forty-three counties in Shanxi Province were taken as cases to test the model. Results show that ACA reduces the number of indicators. Form 30 indicators of the index system of water-saving irrigation technology, 12 indicators were selected out for wheat, 16 indicators for corn, 17 indicators for soybean and cotton. The model was used to optimize water-saving irrigation for wheat, corn, soybean and cotton fields respectively, and results were basically consistent with local conditions. The proposed model could provide scientific basis for decision-makers, and has great practical significance in selecting suitable water-saving irrigation technology for planning and designing irrigation projects.

In order to reduce the influences of the operation temperature of Distributed Feedback (DFB) laser on its output wavelength and optical power, a strong robustness DFB laser temperature control system was developed using Ziegler-Nichols Proportion-Integral-Derivative (PID) control algorithm. Using this temperature control system, an experiment was performed on a DFB laser with a center wavelength of 1.742 μm, which was manufactured by the Institute of Semiconductor, CAS. The results indicate that the temperature accuracy of ±0.05 ℃ and control range of 5 ℃-60 ℃ can be achieved by this system respectively. The tested DFB laser works stably without center wavelength shift during long term (220 min) operation. Therefore, this temperature control system provides superior performance that guarantees the application of the DFB laser in infrared gas detection.

In order to get the absolute diffraction efficiency of transmission gratings, a soft X-ray monochromatic system is built up. The system is composed of a concave reflective grating monochromator and a laser plasma source. Three spectral lines of 30.38 nm, 53.70 nm and 58.43 nm from Helium hollow cathode source are used to calibrate the wavelength of the transmission grating measuring system. A black illumination Charge-Couple Device (CCD) is used as reception element to experimentally calibrate the diffraction efficiency of a golden self-supporting transmission grating at several energy points between 85 and 900 eV. The calibration results for the first-order relative efficiency and zero-order absolute efficiency are obtained. The diffraction efficiency of this grating in soft X-ray band is simulated using quasi trapezoidal grating line model. The simulated first-order relative diffraction efficiency is basically consistent with experimental calibration. Finally, the measured zero-order absolute efficiency and simulated first-order relative efficiency are used to fit the curve of the first-order absolute efficiency of the grating.

An appropriate data pre-processing method is needed to improve the modeling precision for oil yield analysis of oil shale in situ by the near infrared spectroscopy(NIRS)analysis technique. With nature and proportioning samples, the spectra preprocessing and Partial Least Square (PLS) modeling experiments were carried out by two ways (spectra and matrix) of data preprocessing. Results show that the appropriate spectra preprocessing (including first derivative, auto-scaling, mean centralization and some combined methods, such as adding mean centralization after smoothing, denoising and de-baseline, respectively) can enhance the spectrum identity of the same sample. Three combined data preprocessing methods (including combined mean centralization or first derivative spectra preprocessing with auto-scaling matrix preprocessing) can improve the precision of NIRS analysis of PLS model for oil yield from proportioning samples.

In the time domain electromagnetic detection system, emission current overshoot can cause receiving blind area. In order to solve this problem, AC equivalent models of current overshoot produce and inhabitation are established. The circuit theory is applied to the analysis of the proposed model, and a method of accelerating resistance increase and damping of overshoot current oscillation is put forward. As a result the time domain electromagnetic detection transmitter can be improved. Experiments show that the improved system is still stable and reliable; early signal of wave is advanced by 100 ?s, thus reducing the detection blind area. The proposed method provides hardware premise for current waveform features of time domain electromagnetic detection. It also provides reference for the design of electromagnetic instruments on unmanned aerial vehicles.

In order to reduce the switching loss of ICPT convertor and simplify the soft-switching control circuit, the push-pull transformer was introduced by connecting the transformer coil and the compensation capacitor on the output of the push-pull transformer. The converter can achieve zero-voltage turn-on by making the transformer coil and the compensation capacitor resonant network to present inductive. It can also achieve zero-voltage turn-off by adding capacitors between the drain and source of the switch. Therefore, there is no need to increase the voltage zero-crossing detection and the corresponding control circuits to achieve soft-switching. The working principle of the soft-switching was analyzed. The zero-voltage conditions and the boundary conditions of the number of turns of the transformer were given. Simulation and experiments show that the convertor can operate in soft switching conditions, with the advantages of simple drive and high transmission efficiency.

To solve the problem of simultaneous localization and mapping by rescue robot in line feature post-disaster environment, an improved dynamic threshold algorithm is proposed based on the split-and-merge principle. Compared with the fixed threshold method, the proposed method improves the accuracy of straight line fitting. The parameters of matching conditions and state matrix are deduced. Mismatching problem can be solved by calculating and comparing the distance between initial point and end point, thus improving the accuracy of line feature matching. On this basis, global location calculation is carried out and least square filter is employed to improve accuracy of location. The simulation and experiment demonstrate that the proposed method has high accuracy to reflect the actual terrain environment and the ability of real-time mapping.

Composed problems are one kind of random Constraint Satisfaction Problems (CSPs). When solving such problems, existing heuristics always bring out excessive nodes, thus seriously affecting the solving efficiency. Based on the structural property of the Composed problems, we develop a new variable ordering heuristic, named boundary heuristic. First, we compose the concept of boundary variables. Then we present the method to recognize boundary variables in Composed problems. During search process, boundary variables have higher priorities to be instantiated than common variables. We propose two strategies to enforce the new heuristic when it is used to solve problems. Experimental results show that the efficiency of the boundary heuristic outperforms the existing ones by large margins.

As the transitions of states for software program have inequality constraints, a formal description, named linear semi-algebraic transition system, is proposed for program. With using characteristic set, a novel notion of bisimulation is established to simplify the structure of programs on the basis of common zeros of linear semi-algebraic systems. However, in applications, most of experimental data only have approximate values within an given error range. Therefore, the exact equivalence relations lack fault tolerance and flexibility. To obtain more appropriate relations for software programs, Singular Value Decomposition (SVD) is used to approximately process the systems and achieve the notion and calculation of approximate bisimulation. By symbolic and numerical hybrid computation, the error is computed and controlled. Finally, the approximate system is used to replace the original complex system. A case analysis of concurrent communication program shows that our approach is reasonable and effective for simplifying software program design.

One of the main objectives in the analysis of microarray data is the identification of Differentially Expressed Genes (DEGs) under different experiment conditions. A main approach for such analysis is to calculate a statistical value for each gene, and then rank the genes in accordance with their statistical values. A large ranking value is evidence of a differential expression. Inevitably, different methods generally produce different gene rankings, and the performance of each method depends on its evaluation metric, the dataset and data preprocessing method. A disadvantage shared by existing methods is that some top ranked genes, which are falsely detected as DEGs, tend to exhibit lower expression levels. Here, we present a novel technique named Matrix Rank Product (MRP) for identifying differentially expressed genes that originate from a simple statistical rank model. The algorithm can directly deal with the raw data of the microarray. As a result it can eliminate the interference of different data preprocessing methods. Meanwhile, the new technique is designed for accurate gene ranking by calculating the microarray data matrix of overall sorting.

Based on the introduction of Glowworm Swarm Optimization (GSO) and Back-Propagation (BP) neural network, a target threat assessment model is proposed and its algorithm is developed. This model is based on GSO optimized BP network (GSOBP). In GSOBP, GSO is employed to simultaneously optimize the initial weights and thresholds of the BP neural network. Target threat database is adopted to test the performance of GSOBP in target threat prediction. The performance of GSOBP is compared with that of normal BP neural network and Particle Swarm Optimization and Support Vector Machines (PSO_SVM). Experiment results show that target threat prediction accuracy by GSOBP is higher than that by normal BP or by PSO_SVM.

In order to solve the sample impoverishment phenomenon, the Invasive Weed Optimization method was introduced into Sample Inspection Report (IWOSIR) of generic particle filter. This method incorporates the newest observations into the sampling process, enabling the particles to reproduce dynamically in the nearby space with their own fitness, and optimizes the particle population with optimal weights. Though IWOSIR, particles are moved towards the regions where they have larger values of posterior density. As a result, the approach relieves the effect caused by sample impoverishment of particle filter though ameliorating the diversity of sample set. Simulation results demonstrate that IWOSIR has higher estimation accuracy and operational efficiency.

Network coding is analyzed from time sequence in terms of the parity check matrix. As a consequence, the sequence matrix description is given for encoding and decoding of network codes. Meanwhile, the deterministic linear network coding, random linear network coding and convolutional network coding are unified in this framework, which provides deep understanding of network coding. Based on this analysis, a decoding approach is proposed with polynomial-time complexity, the feasibility of the approach is analyzed. Besides, the decoding principle and some properties of the convolutional network coding are discussed. Finally, a case study illustrates the decoding procedure of network coding using sequence matrix description.

A routing protocol SECMR based on synthetical estimation of contact metrics and forwarding constraint is proposed when there exist several social sub-regions constituting delay tolerant social network. By defining the social state parameter of nodes, the proposed SECMR characterizes the motion of nodes using movement model, which combines Integrated Product Meta Model (IPMM) and Random Waypoint (RWP) model, while intra-region forward constrained factor SOC_CST is introduced to avoid data congestion in active nodes. Simulation shows that SECMR strongly accommodates social Delay Tolerant Network (DTN). Compared to the performances of Prophet and MEED, massage delivery is improved by 15% and 20%, average transmission delay is improved by 9% and 12% with a small increase in resource overhead.

In order to improve the accuracy of reconstructing Bandpass Prolate Spherical Wave Function (BPSWF), a fact convergent reconstruction algorithm is developed. According to the sampling theorem and the fact that bandpass signal can be recovered from its well sampling signal with a suitable bandpass filter, a general construction formula is given. Main factors inducing the errors of the reconstruction algorithm are analyzed, and the basic idea of using more convergent base functions to improve the reconstruction accuracy is proposed. A group of fast convergent base is given by design in frequency domain; then, the fast convergent algorithm of reconstructing BPSWF is proposed. Theoretical analysis and simulation results demonstrate that the new reconstruction algorithm is fit for recovering BPSWF of low frequency. Compared with the reconstruction method with sinc base function, the proposed algorithm can achieve higher reconstruction accuracy, which makes the recovered BPSWF get better orthogonality and energy concentration characteristics.

Target based Cooperative Spectrum Sensing (CSS) scheme with best relay (P_e-BRCS) was proposed. In this scheme the best relay was selected by minimizing the reporting error probability to improve the sensing performance. Furthermore, to reduce the computational complexity, a simplified CSS scheme with best relay (simplified-BRCS) was proposed. Results of numerical simulation show that, comparing with traditional SINR-BRCS, the reduced reporting error probability can be achieved effectively by the proposed Pe-BRCS with little computational complexity. The performance obtained by the simplified-BRCS scheme is similar to that of Pe-BRCS but without increase in computational complexity.

To achieve the joint estimation of Direction-of-Arrival (DOA) and Doppler frequency in array signal processing, the artificial bee colony theory and algorithm are applied to optimize the likelihood function. First, the extended observability matrix containing the angle-frequency parameters is established using state-space model. Then, the likelihood function containing the extended observability matrix is used to convert the problem from parameter estimation to nonlinear multidimensional function optimization. Finally, artificial bee colony algorithm is applied to optimize the process of finding the solutions of the maximum likelihood estimator in order to estimate the DOA and Doppler frequency. The proposed method reserves the asymptotic unbiased estimation of the maximum likelihood estimator and reduces the computational burden to calculate the solution of the likelihood function. Besides, the estimated parameters can be paired automatically.

The existing video compression standard possesses certain limitations on the support and expansion of the Multi-View video Coding (MVC). To solve this problem, combined with the multi-dimensional vector matrix theory, a whole set of compression encoding schemes on single-view colorful video stream proposed in previous studies was expanded and applied in eight-view video coding. This multi-view video coding includes multi-dimensional sub-blocking, multi-dimensional reorganization, multi-dimensional vector DCT orthogonal transform, multi-dimensional quantization, differential encoding, multi-dimensional scanning and run length encoding. Analysis of the experiment data verifies the feasibility of the whole scheme on multi-view coding. This research lays a foundation for further study of both multi-dimensional vector matrix and multi-view coding.

We propose an algorithm for optimal homography matrix in video mosaic using double thread. First, the characteristic points are extracted using Speeded Up Robust Feature (SURF) method. Second, the corresponding characteristic points are matched using Nearest Neighbor (NN) and Random Sample Consensus (RANSAC) methods. Third, the optimal homography matrix of first k video frames is calculated using the normal covariance correlation of overlapped area between corresponding frames. The optimal homography matrix is taken as the mapping matrix of the succeeding frames. Meanwhile, the KLT algorithm is used for feature dynamic tracking of the frames succeeding k+1 frame. If the change of amount of the tracked feature points exceeds the given threshold, the current optimal homgoraphy matrix needs re-optimization and transformation; the pairs of feature points of k video frames are recalculated and matched to obtain the new optimal homgoraphy matrix. Image matching tests in traffic surveillance demonstrate that the average matching speed of the proposed algorithm is less than 100 ms; good matching quality is achieved for images with variable rotation, scaling, visual angle and illumination; parameter estimation is accurate and high speed with less computation, which meets the requirements of real-time and high accurate matching of the system.

A new method of face recognition based on relevance vector machine was developed. After the wavelet decomposition and Principal Component Analysis (PCA) transformation, the relevance vectors from sample training constitute a "hyperplane" as the differences in the classification of the samples by machine learning algorithm. The "one against one" method is used to achieve multi-class pattern recognition. Compared with former methods, a large number of simulation results of recognition of noisy objects show that the new method is not sensitive to image noise, with high recognition accuracy and strong robustness.

This article puts forward a new method of image fusion based on interpolation processing of Directionlet and compression sensing theory to highlight the details of regional characteristics and improve the image visual effect. After interpolation Directionlet transform decomposition, the low-frequency component coefficients are fused using the regional characteristics and Verbal Immediacy; and the high-frequency component coefficients are fused using the improved sparse Hadamard matrix observation weight fusion method. Then, the fused high-frequency component coefficients are reconstructed. Finally, the image is reconstructed by the inverse interpolation Directionlet transform. Experiment results show that this fusion algorithm can get image with clear object and background, and it has low computational complexity and good fusion effect.

This paper analyzes various image restoration methods for different image degradation models and discusses how to apply deconvolution to match suitable blur model first. Then, on these bases, a new method with denoising preprocessing is proposed to realize underwater image restoration without prior information. The proposed method considers the Radon transform and Fourier transform to estimate the fuzzy angle and distance in the Point Spread Function (PSF), which is simpler and more precise than other estimation methods. Simulation results show that this denoising preprocessing could inhibit the ill-posed problem of image restoration process, thus the proposed method can greatly improve the image quality and definition.

In order to solve the target tracking problem when the statistic characteristics of the system are unknown, an adaptive unscented particle filter algorithm is proposed. This algorithm estimates and corrects the statistic characteristics of the unknown system noise in real-time using improved Sage-Husa estimator. Combining with unscented Kalman filter, the algorithm produces the optimal proposal distribution function. This method effectively reduces the estimation error and improves the anti-noise ability of the system. Theoretical analysis and experiments show that the new method can significantly improve the accuracy and stability of target tracking when the statistic characteristics of the system are unknown.

Before determining the carrier-phase ambiguity resolution of BeiDou satellite in precision positioning, carrier-phase cycle slips need to be detected and repaired. Due to the influences of different errors, the carrier observation equation is nonlinear and non-Gaussian, so the method of linear filtering has its limitations to detect cycle slips. In this article, nonlinear and non-Gaussian adaptive particle filtering is employed to detect cycle slips. The detection model is built by observation equation based on analyzing BeiDou cycle slip detection principle and computation process. The proposed method can meet the demand of BeiDou cycle slip detection and precision positioning using measured data. Compared with Kalman filtering method, the adaptive particle filtering cycle slip detection model is characterized by simple initialization, high accuracy and fast convergent speed.