Abstract:In order to decrease the clamping force and increase the transmission efficiency of CVT variator system， the relationship between the slip rate and comprehensive efficiency of variator system was examined by using the experimental identification method. Considering the coupling effect of slip on the transmission loss and oil pump power consumption， a model for testing the comprehensive efficiency of variator system was constructed. Based on a self-developed CVT variator system， the tests of transmission efficiency to slip rate and oil pump power consumption were designed， the variation characteristics of the comprehensive efficiency of variator system considering the slip rate under different speed ratio， input torque and input speed were analyzed， and the maximum efficiency slip line and maximum efficiency safety factor of the variator system were derived according to the test results. The results demonstrate that， with the increase of the slip rate， the peak comprehensive efficiency slip point of variator system occurs， which is located between the peak transmission efficiency slip point and friction coefficient saturation slip point， and mainly affected by the speed ratio. The influence of input torque and input speed can be neglected. In the whole speed ratio range， the maximum efficiency slip line of variator system is located between 0.98%~3.16%， the corresponding comprehensive efficiency is between 0.878~0.915， and the maximum efficiency safety factor is between 1.09~1.16， which is reduced by 10.8%~16.2% when compared with the traditional clamping force safety factor. The results of the study can provide a clear control target for CVT slip control and also provide a lower reference boundary for reducing the safety factor of CVT clamping force.

Abstract:In order to improve the shifting smoothness and to reduce the shifting time of the two-speed Automated Manual Transmission（AMT） equipped on the pure electric vehicle， the mathematical model of the powertrain system was developed， and the comprehensive coordinated control strategies in which the motor participates in the shifting process were designed. The motor torque control law of torque phase was derived from the maximum allowable shift impact，and the PID and finite state switching control strategy were adopted to regulate motor speed during the inertia phase. Additionally， a test rig of the pure electric vehicle transmission system was built for the up-shift and down-shift test. The simulation results show that the up-shift time of 0~100 km/h full acceleration is 0.5 s， the longitudinal impact is within 8.0 m/s3， and the up-shift and down-shift time in the NEDC urban conditions are both within 0.6 s， and the maximum impact does not exceed 7.8 m/s3. The test results show that the up-shift and down-shift time is 0.6 s and 0.8 s，respectively，which is in the condition that the drive motor under fixed speed control， and the output shaft speed changes smoothly. Because the shift time of the traditional AMT vehicle is 0.8~1.0 s， the above results show that the comprehensive coordination shift control strategy can achieve fast and smooth shifting.

Abstract:In the pedestrian-vehicle collision accident, pedestrian may suffer head injury not only from the primary impact with the vehicle but also from the secondary impact with the ground. In order to comprehensively analyze the influence of vehicle front-end structure parameters on pedestrian kinematics and head injury in primary and secondary impacts，in this paper, a series of multi-body models of pedestrian-vehicle collisions were constructed by utilizing the MADYMO code, with adjusting parameters of windshield angle（WA），bonnet angle（BA），bonnet length （BL），bonnet leading edge height（BLEH） and ground clearance（GC）. The influence significance of vehicle parameters on the pedestrian kinematics and injury was analyzed by using multiple linear regression analysis. The results indicate that BLEH is the most significant influence factor on head HIC value（car）（p<0.01) and pedestrian rotation angles in the primary and secondary impact，respectively.The influence of BA on the head angle acceleration （car） is more significant than that of BLEH. WA, BL and GC show a less influence on the kinematics and injuries of pedestrian. The lower BLEH provides better protection of the head in the primary collision, but it can easily lead to the head to collide with the ground firstly in the secondary impact, which increases the risk of head and neck injury. A quantitative analysis of the effect of vehicle front-end structure parameters on the primary and secondary impact was also conducted in the study, and the results can provide some reference for vehicle safety design.

Abstract:Rearview mirror is an important source of aerodynamic noise in the front window area of the car. It is of great significance to study the characteristics of the wake of rearview mirror to understand the mechanism of aerodynamic noise of rearview mirror. The wind tunnel test platform was built to measure the time-averaged and instantaneous wake of the rearview mirror bluff body by surface oil streamline, hot wire anemometer, and Particle Image Velocimetry（PIV）. Proper Orthogonal Decomposition（POD） was used to identify the large-scale coherent structures in the wake of the rearview mirror. The experimental results show that, due to the finite length feature, the wake of the rearview mirror is similar to that of the finite-length cylinder. Alternating vortex shedding exists in the wake, whose characteristic frequency is less than 0.2. Under the influence of the downwash airflow at the top end, the characteristic frequency of the vortex shedding is unevenly distributed in the streamwise direction and the spanwise direction. The POD analysis results verify the existence of alternating vortex shedding, which leads to the streamwise and spanwise oscillation of the wake recirculation zone. Therefore, the wake of the rearview mirror is similar to that of the finite length cylinder, and the wake vortex shedding is mainly affected by the finite length characteristics.

Abstract:In order to facilitate the quantitative analysis of hydrogen-consumption impact factors and hydrogen-consumption potential for fuel cell vehicle, the average integrated transmission efficiency was first defined and a theoretical model of hydrogen-consumption calculation and analysis was proposed, on the basis of the internal energy flow relationship of the fuel cell vehicle during driving. By analyzing the various impact factors of hydrogen-consumption, the theoretical hydrogen-consumption-increment model was further obtained. Finally, based on the fuel cell vehicle model provided by the advisor, the effects of rolling resistance coefficient, mechanical efficiency of the drive train, motor efficiency and fuel cell efficiency on the economics of the vehicle were quantified through theoretical and simulation analysis. In addition, based on the possible-limit state that each influencing factor may reach in the future, the hydrogen consumption potential obtained from the simulation is 0.6 kg/100 km. The quantitative analysis of different factors and the determination of hydrogen consumption potential not only provide critical reference for future research of the fuel cell vehicle powertrain, but also point out the optimization direction for component selection and parameter calibration during the development of actual vehicle.

Abstract:The heat dissipation of lithium-ion battery pack has been an important factor affecting the battery life and traffic safety of electromobile. In order to explore the effect of different cooling pipeline design on the heat dissipation of lithium-ion battery pack， the surface temperature of a single lithium-ion battery under different conditions was studied by means of numerical calculation，and compared with the test results，the correctness of the simulation method was verified. Under the 27 ℃， the heat dissipation effects of eight different cooling structures were compared and analyzed， finding that the average temperature of 8th structure is 31.62 ℃ and the standard deviation is 0.83， which has the best cooling effect； the design of bidirectional flow， the inlet position ，the shunt situation of branch pipe， and the contact area between the cooling pipe and the battery pack will have different influences on the heat dissipation performance of the battery pack. Therefore， these factors should be taken into consideration in the design of the heat dissipation structure of lithium- ion battery pack.

Abstract:The application of Tailor Rolled Blanks （TRB） is one of the most effective ways to achieve automobile lightening， however， it's difficult to join the materials with the variable thickness. The thickness of B340LA TRB ranges from 1.2 mm to 2.0 mm and fluctuates in both length and width directions. And its hardness has a maximal deviation of 20% at different thickness. The special welding fixture was designed based on characteristics of TRB， and step variable laser power was proposed. By a contrastive analysis of two laser welding tests， including constant laser power and step variable laser power， the results show that there are huge differences in the weld appearance and penetration status along the welding direction under the conditions of constant laser power（900~1 400 W），and the defect of surface depression occurs in a portion of weld. By step variable laser power with 1 000 W， 1 100 W and 1 300 W，approximately consistent weld appearance and penetration status were obtained in the direction of thickness from 1.2 mm to 2.0 mm， and the hardness deviation in fusion zone of weld is less than 9%. B340LA TRB can be joined more effectively by step variable laser power than by constant laser powers.

Abstract:In view of the existing problem in the wavelet analysis of acoustic emission signals in wear state of diamond grinding wheel， because engineering ceramics partially stabilized zirconia grinding acoustic emission signals have nonlinear and nonstationary characteristics， using empirical mode decomposition method the acoustic emission signals were decomposed into several stationary intrinsic mode functions and then the root mean squares， variances and energy coefficients were extracted. When the wear state of diamond grinding wheel changes from mild wear to severe wear， the root mean squares（IMFrms） and variances（IMFvar） of the intrinsic mode function increase， and the energy coefficients（IMFpe） change significantly. As the input parameter of the least squares support vector machine， the wear state of diamond grinding wheel was successfully monitored.

Abstract:Perturbation method and Poincaré mapping method were used to derive the generalized Melnikov function of the periodic solution for a two-degree-of-freedom vibro-impact system with cubic non-linearity and external excitations. By using the Melnikov′s method, the characteristics of periodic motions with double-impact of the 2-dof system were studied, and the existence condition of period-2 motions with double-impact was determined as a critical curve in the parameter domain. The results of numerical simulations show that the regions below the critical curve are the period-2 motions with double-impact, the upper regions of the critical curve are not period-2 motions with double-impact；Meanwhile，increasing the force amplitude and keeping the other parameters unchanged, the motion state of the system changes from multi-period motions with multi-impact to period-2 motions with double-impact, while increasing the system restitution coefficient and keeping the other parameters unchanged, the motion state of the system changes from period-2 motions with double-impact to multi-period motions with multi-impact.

Abstract:To solve the serial dependency problem that came up with CABAC in the next generation of video coding standard called High Efficiency Video Coding（HEVC），this work proposed a low redundancy full system CABAC decoder solution. By adopting the dynamic codebook and parallel computing design, this work improved the clock efficiency and could satisfy the HEVC Level4.1 Main Tier and decode a 1080HD@60fps video with a clock of 40MHz.

Abstract:Power is an important factor that restricts the design of Integrated Circuit (IC). The power origin in CMOS IC，the aim of power estimation in IC design，the estimation method，and the power model were analyzed. The characteristic of analog integrated circuit and the corresponding way for power estimation were studied. In order to distribute the power at system level, the construction of charge pump PLL，the principle of every module and the contribution to power were also studied. The power estimation model of charge pump PLL with ring oscillator was proposed at system level. The relative error is less than 22% when compared with the actual measurement. This model is easy to be integrated into design tools and can give guidance on power consumption for PLL at system level to improve the quality of Integrated Circuit design.

Abstract:Based on the development requirements of low-cost and low-power Internet SoC chip, an image acquisition and control system is designed and implemented based on open source RISC_V instruction set architecture SoC chip and SMIC 55nm CMOS process, which uses the low-power open source processor RI5CY SoC chip as a platform and combined with the Low-Voltage CMOS Image Sensor OV7725 with the integrated DSP and A/D conversion chip inside. This paper introduces the block diagram of the image acquisition and control system, and describes a design of image acquisition controller based on AHB bus in detail. An improved asynchronous FIFO with low power and small area is used to realize the synchronous design of different clock domains. According to the results of Modelsim simulation, DC synthesis and FPGA verification shows that the system realizes the video image data acquisition and transmission，which operation flow is simple and easy to debug, and the maximum of the data transmission bandwidth come up to 37MB/s. The chip total area is 3250 μm x 3648 μm and power consumption is only 24.419 mW.

Abstract:Measurement accuracy of γ-ray backscatter ash analyzer is greatly affected by the changes in the composition of the coal ash. In current ash analyzer, calibration is underway anew when measurement error is comparatively large, with a narrow measurement range, low accuracy and complicated operation. In this paper, the impact of changes in the composition of the ash on the accuracy of ash analyzer was analyzed from the working principle of γ-ray backscatter ash analyzer. A calibration model based on the dual low energy γ ray measurement was established，which can improve the measurement accuracy and can be achieved automatically on software. The calibration model was then used on a γ-ray backscatter ash analyzer, and the test results show that the existing deficiencies are overcome with an measurement error of less than 1%，meeting the required measurement accuracy of EJ/T1078-1998 γ- ray coal ash analyzer.

Abstract:When deep learning is used to deal with the computer vision tasks, under little number of new task data, the pre-trained model weight based on a very large data is trained as an initial weight to get better generalization ability. At this point, former explanations are based on the intuitive analysis and lack of reasonable mathematical methods. In this paper, deep neural network, which trains on internal layers with fixed structure，changed into internal transfer ability in deep neural network. The changes of learning process are formalized into a mathematical expression. Considering the influence of the data set on the training process, the information geometric analysis method is used to determine the metrics and connections over manifolds of different data sets, which can realize the embedding mapping between different data sets. At the same time, the change of parameter space is also put into a manifold space to explore its common influence on learning process. Finally, a mathematical explanation is provided for the internal transfer phenomenon. Meanwhile, after the analysis and experiments, the process of internal transfer is identified as a change which can make the network search for optimal search in a wider space. Therefore, the model can obtain a relative better solution in learning process.

Abstract:A new method was presented for computing the tight-fitting enclosing minimum volume oriented bounding boxes for constructing the model of complex object point sets in three dimensions. The relationship between the convex hull and its minimum volume oriented bounding box with the smallest volume was analyzed，and four kinds of edge contact types were summarized. The optimal box orientations are uniquely determined by combinations of edges in the convex hull of the input point set. Empirical evidence shows that this process always yields the globally minimum bounding box by volume feature, which concludes that this method provides a good simulation.

Abstract:To cope with the problem of object tracking failure in the challenging environment, a target tracking method based on multi-correlation filter combination was proposed. Firstly, two kernelized correlation filters（KCF） based on color name（CN） features and histogram of oriented gradient（HOG） features, respectively, integrated the map information through adaptive fusion method, and were used to determine the prediction position of the target. Then, through the multi-scale sampling based on the target region, CN-HOG compositive feature was extracted to construct a scale correlation filter to obtain the optimal scale of target. Finally, the adaptive updating strategy of the model was designed to determine whether the model was updated in the current frame through determining whether the target was occluded. The proposed algorithm and 6 state-of-the-art methods were tested on 50 video sequences. The experiment results indicate that the proposed algorithm gains the best precision and success rate in the challenging environment, it can effectively deal with the problem of object occlusion and scale change, and it has a fast tracking speed.

Abstract:To accurately and reliably estimate the solar irradiance, a random forest algorithm was proposed based on the Bayesian model combination for solar irradiance prediction. Firstly, the K-means clustering and K-fold cross validation were introduced to generate multiple training subsets so as to increase the diversity of training subsets and to ensure uniform sampling. Secondly, the random forests were defined as base learners to establish an ensemble learning prediction model，with each training subset being used to train the corresponding individual random forest. Then, according to the prediction performance of each individual random forest on the verification set, the Bayesian model combination algorithm was applied to formulate the combination strategy. The prediction values of individual random forest on the test set were fused to the final output through the model combination strategy. Finally, the proposed method was applied to solve the solar irradiance prediction problem. Simulation experiments were carried out by measured meteorological data. Other four kinds of prediction methods were also introduced to establish the contrast experiments，and the accuracy and reliability of the proposed method in the solar irradiance prediction were verified by comparison results.

Abstract:Entities mostly exist in heterogeneous forms in Internet of Things (IoT)， so solving heterogeneous issue is a vital task for developing IoT. On the basis of analysis on the characteristics of entities and their uploaded information in IoT， an agent-based heterogeneous entities relation service model was proposed. The functions of model include entity information data processing and entity service logic construction， and through using the new inter-objects dynamic relation calculation formula proposed in the paper，the heterogeneous data was transformed into their relation closeness（values），and the agent service sequence was generated based on it. Combining with the bipartite graph theory， the agent service logic was realized in IoT environment， which overcomes the heterogeneity that entities are unable to communicate with each other properly. In the model， entity interactions were completely replaced with agent interactions，and these agents were governed under the established service logic. Finally， the simulation experiment was performed on the model. And the result shows that not only can user's service requirements in IoT environment be met but also it is better than the traditional ontologies and semantics technology to deal with heterogeneous issues in the data updating， service quality and conflict handling.