The tire low-frequency handling model and the high-frequency durability model were unified， and the UniTire tire model including in-plane dynamic characteristics was proposed. The steady-state UniTire is the zero-frequency output of the tire dynamic system. When the frequency of the tire motion input signal is infinite， the output of the wheel center force is zero. Using these as boundary condition， the transfer characteristics of the tire dynamic system are established. Considering that in the simulation of the vehicle virtual proving ground （VPG）， the load history of the wheel center is mainly affected by the rigid ring modes of the tire， that is， the in-plane frequency is less than 100 Hz， and the out-of-plane frequency is less than 60 Hz. Higher frequency components contribute to the load analysis of the wheel center indistinctively， so a second-order transfer model that satisfies the boundary conditions is established. Finally， using typical tire dynamic test data， the tire model parameter identification and test verification are carried out. The comparison results show that the proposed model can accurately express the in-plane tire dynamic characteristics.

The Converge was used to study the spray characteristics of porous injectors under different fuel temperature conditions， and the influence on the spray collapse degree was studied by changing the number of nozzles， ambient pressure and temperature， oil beam injection direction and radial spacing of adjacent nozzles. The results show that with the increase of fuel temperature， the degree of flashing phenomenon is more violent， the spray cone angle and spray width increase， the spray center is difficult to exchange with the external gas， so that the low pressure condition of the spray center is difficult to improve， and the oil beam gathers to the middle under the action of pressure difference， and finally evolves into spray collapse. By changing the ambient pressure and temperature， it can be found that after appropriately increasing the environmental pressure， the spray collapse phenomenon will become violent， and increasing the ambient temperature has no obvious effect on the degree of spray collapse； Changing the oil beam spacing by changing the number of nozzle holes， the direction of the oil beam and the radial spacing of the spray holes has a significant effect on the degree of spray collapse.

Aiming at the problem of vehicle steering nonlinear equilibrium state and its stability determination， a new method for solving the equilibrium state of vehicle steering nonlinear two degree of freedom model and its stability criterion was proposed. Firstly， the nonlinear two degree of freedom model of vehicle steering was established， and the simplified magic formula tire model was used to describe the nonlinear tire lateral force. Then， the particle swarm optimization algorithm for solving the nonlinear equilibrium state of the corresponding model was established， and the stability criterion of the equilibrium state was deduced through theoretical analysis. Finally， three driving conditions with different front wheel steer angle， road adhesion coefficient and vehicle speed were designed， and the particle swarm optimization algorithm and the quasi Newton method were used to solve the steering nonlinear equilibrium state under the three driving conditions and to judge the stability of three kinds of driving conditions respectively. The results show that the proposed method can accurately solve the equilibrium state of the vehicle steering nonlinear two degree of freedom model， the solution effect is better than the quasi Newton method， and the stability criterion is suitable for determining the stability of the equilibrium state.

Aiming at the problems of non-linearity and uncertainty disturbance in slip rate control in the emergency braking process of automobiles， and the difference in the bandwidth characteristics of the braking system， the slip rate controller cannot be effectively applied to the actual vehicle. An implementation based on active disturbance rejection control method was proposed for setting the bandwidth parameters of the converter. First， the vehicle model and the ABS model were established. Secondly， an active disturbance rejection control with differential tracker was designed， state observer and error feedback controller were extended. Then， based on the ABS braking bandwidth parameters， the active disturbance rejection control parameters were adjusted. Finally， the slip rate control simulation analysis was carried out. The feasibility and effectiveness of the active disturbance rejection control method for ABS slip rate are verified. The results show that the method has clear physical meaning， simple parameter setting， strong portability， and can realize precise control of slip rate.

Aiming at the problem of electric bus out of control caused by driver's wrong stepping on the pedal， natural driving test and simulated driving test of wrong stepping on the pedal were carried out. The pedal misoperation identification methods were established based on One-Class SVM and iForest respectively. The verification results show that the four parameters could be used to reflect pedal misoperation behavior， vehicle speed， motor torque， accelerator pedal opening and brake pedal opening. The recognition accuracy of one class SVM and iForest identification methods are 94.9% and 99.5% respectively， but iForest identification method has outstanding advantages in the recognition accuracy of pedal misoperation. The iForest method can be used to identify the running data of electric bus， which can provide theoretical support for reducing the accidents related to abnormal braking.

Due to its poor rigidity， aluminum alloy thin-walled parts are easily affected by the spindle speed， axial cutting depth， feed speed and radial cutting depth， resulting in chatter in the milling process. In order to solve this problem， an active flutter suppression method for milling aluminum alloy thin-walled parts was designed. The milling dynamics model based on cylindrical spiral end milling cutter was established， the cutting force changes in x and y directions were analyzed， and the mechanical trajectories of cutter teeth during cutting were calculated. The wavelet neural network proportional integral derivative control algorithm was used to realize the active suppression of the chatter in milling parts. The experimental results show that the method can achieve stable milling under high tool speed and axial cutting depth， and the cutting force after inhibition is very consistent with the optimal cutting force， and the vibration range after inhibition is controlled between -5 μm and 5 μm， which can reduce the deformation and damage in the processing， thereby improving the production efficiency and product life.

Flatness is an important shape deviation. A flatness error measurement method was proposed based on line structured light vision technology for flatness measurement in the field of mechanical component manufacturing and processing. Firstly， the images of light strips were collected， and the spatial coordinates of the scanning points were obtained by the corresponding light plane equation for each position. Secondly， the evaluation methods for flatness errors in national standards was analyzed and a measurement algorithm for flatness errors was established based on geometric constraints. Finally， through the proposed method， the evaluation base plane and flatness errors were calculated using the spatial coordinates of the scanning points. In the experiment， the positioning surfaces of the insert molds are selected as the measured planes， and the measurement results obtained by visual measurement are compared with those obtained by the contact measurement method. The measurement error is less than 20 μm. The experimental results show that the measurement method proposed is feasible and improves the measurement efficiency of flatness error.

The abrasive flow precision machining method was used to deburr the cross hole of the valve sleeve. Through the large eddy simulation method and the reasonable sub-grid model， the numerical simulation analysis was carried out for the valve sleeve workpiece under the condition of side hole outflow. Under the condition of different flow velocities and viscosity， the effects of fluid strain rate， velocity vector， and wall shear force on the fillet radius of cross-hole edge， burr removal， and radial hole wall processing quality were analyzed. The experimental results show that the lower the fluid viscosity and the higher the speed， the larger the fillet radius of the edge of the cross hole， the higher the viscosity and the speed， and the higher the processing quality of the radial hole wall. To ensure that the edge of the cross hole has a small fillet， the low-speed and high viscosity fluid should be selected for the precision processing of the cross hole of the valve sleeve.

To solve the inverse kinematics of humanoid redundant manipulator， an optimized method was proposed. The analytical solution of the inverse kinematics was established by introducing the arm angle parameters， and the feasible range of the arm angle under different forms of analytical solutions was further analyzed according to the joint limit. In addition， an objective function considering joint limit avoidance and optimal energy consumption was proposed. Tasks， priority was adjusted through adaptive weight parameters， and the particle swarm optimization algorithm was introduced to obtain the optimal solution. Experimental results show that the algorithm can obtain all solutions for a given arm angle and can select the optimal solution that satisfies the joint limit avoidance and low energy consumption. The proposed method has positive application significance for the trajectory movement of the redundant manipulator.

An anisotropic roll bending fracture prediction method was proposed for 6005A aluminum alloy profile. Von Mises， Hill48 and Yld2004-18p yield criteria were used to describe the anisotropy of aluminum alloy respectively. Model parameters were determined by experiments and crystal finite element simulation. The results of simulation and roll bending test show that the predicted error of Mises yield criterion is more than 15%， Hill48 yield criterion is less than 5% and Yld2004-18p yield criterion is less than 3%. It shows that the finite element model coupled with the anisotropic yield criterion and the continuous damage criterion can predict the fracture behavior of aluminum alloy profiles in roll bending more accurately than the isotropic model.

Finite element simulation method was used to establish the finite element model for the bending process of a “日”-shaped protective beam， and the reliability of model has been verified. The problem of predicting the quality of high-strength steel protective beams has been solved， and the influence of core mold type， profile thickness， and bending forming speed on section deformation has been obtained. The results indicate that the flexible core mold with a self-locking device has the best resistance to cross-sectional deformation. As the thickness of the profile increases， the deformation of the cross-section decreases. As the bending speed increases， the cross-sectional deformation first decreases and then increases. This method is of great significance， which has provided a theoretical basis for predicting the forming quality of protective beam products.

Two redundant operation code-based methods were proposed to solve the energy-efficient driving problem of a high-speed train with steep gradients and speed limits. Based on the necessary conditions for the optimal solution derived from the Pontryagin's maximum principle， the concept of redundant operation code， including redundant operation sequence and switching area， and the generation rules were proposed. Applying the new concept， PMP-LMGA and PMP-PSO algorithms were developed to merge redundant operations and find optimal switching points. The experimental results indicate that the redundant operation code can significantly speed up the calculation in complex scenarios. Total energy consumption can be effectively and stably reduced while meeting various operation rules. Multi-agent parallel computing can further improve solution efficiency.

In the braking process of tractor-trailer-train， due to the different braking sequence of each axle， vehicle instability occurs when braking under special conditions. Since the existing schemes and equipment can not detect the brake timing， an intelligent non-contact dynamic detection scheme for vehicle brake timing is proposed. Based on the principle of monocular vision， the markers affixed to the tire edge are identified. The collected images are processed to solve the tire slip rate. The template matching algorithm based on contour is used to identify the marker region in the image to be matched. In order to verify the effectiveness of the scheme， a real vehicle test is carried out. The test results show that the slip rate solution error is below 4.2%.

In order to explore the heterogeneity of preferences of travel choice and the potential of shared autonomous vehicles to solve the last-mile connection problem of train trips， the last-mile travel mode choice of travelers based on the theory of planned behavior was analyzed. First， a latent variable model was established to obtain travelers' attitudes， subjective norms， perceived behavior control， and behavioral intentions toward autonomous vehicles. Secondly， latent psychological variables into the random parameter Logit model and conduct an empirical analysis was incorporated to study factors of the behavior of the last mile of train trip. Finally， an elasticity analysis was conducted to study the impact of travel time on the travel mode choice. The research results show that the random coefficient Logit model has higher fitness than the traditional multinomial Logit model. Travelers' preferences for travel time are heterogeneous. The travel time coefficient is not a fixed value but follows a normal distribution with a mean of -0.153 and a standard deviation of 0.520. Travelers' attitudes and perceived behavioral control towards autonomous vehicles have a significant impact on travel behavior. The probability of choosingshared autonomous vehicles increases by 0.799% with every 1% reduction in the travel time of autonomous vehicles； The probability of choosing shared autonomous vehicles increases by 1.155 % with every 1% increase in the travel time of shared bicycles.

To clarify passenger cars' longitudinal driving behavior along hairpin curves of mountain roads， the section from Huadiwan to Ningjiazhai of national highway G211 located in Pengshui County， Chongqing， China was selected as the object. The field driving test of passenger cars was carried out to collect trajectory， speed， and acceleration data. Firstly， the characteristics of speed amplitude were analyzed， and the vehicle speed modes in the direction of uphill and downhill were defined. Then， the distribution law of the speed change feature points' position was obtained. Finally， the deceleration and acceleration's amplitude characteristics and influencing factors are studied. The results show that in the curved section， the actual speed is significantly higher than the design speed and speed limit， and the overspeed behavior is very common. The speed pattern of hairpin curves along uphill direction is a ‘decelerate- accelerate’ two-stage speed mode， different from the speed pattern of a standard horizontal curve. The speed pattern for hairpin curves in downhill direction presents three-stage of ‘deceleration-constant speed-acceleration’. The position of speed feature points is affected by sight distance， slope， and deflection angle， which is different from the conventional driving behavior assumption. The speed valley points of the uphill are mainly distributed in the first half of the circular curve. Deceleration endpoints in the downhill direction appear near the SC and MC points； and the acceleration starting point is mainly distributed in the circular curve and the second spiral range. The 85 percentile deceleration rate when entering a hairpin curve on the downhill is 1.25 m/s²， and the one on the uphill is 1.0 m/s². When departing the hairpin curve， the 85th percentile acceleration rate is 0.9 and 0.6 m/s²， respectively， for downhill/uphill curves. Finally， an operating speed model for entering， departing， and staying in the hairpin curves was developed and verify. The research results can provide theoretical and basic data support for the safety evaluation and improvement of complex roads in dangerous mountainous areas.

In order to improve the overall travel efficiency of urban rail transit， an optimization algorithm for urban rail transit operation scheduling based on linear programming was proposed. By extracting the characteristics of rail transit road conditions and passenger flow， a scheduling optimization model that fits the actual situation was established. Linear programming was used to transform the scheduling optimization model into an integrated scheduling optimization model for urban rail transit with significantly reduced infinite discretization probability. The experimental results show that the average delay time after optimization of the proposed method is reduced by 17 min， the travel time of passengers is reduced by 25 min， and the coincidence rate between rail stations and passenger flow demand points is high， indicating that the scheduling effect of the method is good.

In order to study the issue of how to form a more energy-efficient platoon of trucks under the background of vehicle road collaboration， three classic truck models were selected as the research objects and simulation parameters were set with driving speed of 22 m/s and spacing of 4.4 m. Single vehicle （homogeneous） and multiple vehicle （heterogeneous） queues of single， two， and three vehicles were simulated， an ideal simulation calculation domain to analyze the overall energy consumption and energy-saving effect of platoon through XFlow was established. The results showed that the platoon with the highest and lowest energy efficiency rates among the two platoon were the 2-2z and 6-2z fleets， with queue energy efficiency rates of 20.67% and 3.12%， respectively. The teams with the highest and lowest energy efficiency rates among the three platoon are the 2-2z and 6-6-2z fleets， respectively， with queue energy efficiency rates of 31.08% and 3.22%.

The whole road network was divided into sub-networks that meet the requirements of short-term traffic forecast. Three temporal and spatial characteristics are extracted from the short-term traffic volume of the road network， namely time sequence， periodic similarity and spatial sequence， and residual units were added to convolutional neural networks to construct a deep asynchronous network model to realize the short-term traffic flow forecast of the road network. The experimental results show that the root mean square error and average absolute error of the proposed algorithm are small， the trend of traffic forecast value is almost consistent with the trend of real traffic value， and the traffic flow forecast performance is good.

To address the issues of position deviation and inaccurate positioning accuracy in current automatic positioning methods for electric vehicle charging，a machine vision-based automatic positioning method for electric vehicle charging was proposed. The machine vision system in the WSN sensor acquired the driving status of the electric vehicle， extracted environmental images and data information， denoised the charging socket and background using the Hue component and Canny operator in the HSI color model， performed edge detection to determine the center of the coil， calculated the equivalent side length of the charging coil， and accomplished automatic positioning of the electric vehicle through the nonlinear least squares method in machine vision. The experimental results demonstrated that under relative displacements of 3 mm， 5 mm， and 7 mm， the proposed method achieved the same positioning accuracy as the average of multiple positioning attempts， which were 8.24， 8.99， and 8.12. The proposed method demonstrates high positioning accuracy and good positioning performance， as the average positioning accuracy matches the single positioning accuracy. Currently， it is concluded that the proposed method exhibits reliable positioning accuracy.

Aiming at the problem that fire， smoke and other situations at the scene of serious traffic accidents affect the detection equipment to complete the search and rescue of trapped persons， a Convolutional Block Attention Module-Improved Loss-Dual-Discriminator Conditional Generative Based on Convolutional Attention Mechanism Adversarial Network （CBAM-IL-DDCGAN） infrared and visible image fusion method is proposed. Firstly， the decoding network with attention feature fusion module is used to restore and reconstruct the image from space and channel. Secondly， an adaptive weight calculation method based on gradient information is designed. Finally， the test experiment of fused image continuous frames was carried out. Experimental results show that the proposed image fusion algorithm performs well， and achieves a significant improvement of more than 7% in PSNR， SSIM and MSE compared with the traditional algorithm and the generative adversarial network algorithm. These results verify the feasibility and superiority of the fusion algorithm in complex traffic accident rescue.

Durability evaluation of reinforced concrete bridges is a multi-attribute decision-making problem in uncertain environment. According to the characters of uncertain evaluation indicator and its date information， the gray-target assessment model based on improved positive and negative bulls-eye method was proposed， and the concepts of positive and negative bulls-eye distance and projection value， bulls-eye coefficient， positive and negative bulls-eye distance and comprehensive bulls-eye distance were introduced to realize the integrated macro and micro analysis of bridge durability. The model was applied in the durability assessment of nine reinforced concrete highway bridges in Gansu province， the results show that the ranking orders of durability between the nine bridges： Bridge 6#>7#>4#>8#>1#>3#>2#>5#>9#， and the durability of Bridges 6#， 7# belong to Level Ⅱ， the durability of Bridges 1#， 3#， 4# and 8# to Level Ⅲ， and the durability of Bridges2#， 5#， 9# to Level Ⅳ. Through index sensitivity analysis， it is concluded that the absolute average change rate of bridge durability increases linearly with the increase of absolute value of index weight change rate， and there is a certain relationship between index weight and sensitivity， that is， when the index weight change rate is the same， the higher the index weight， the greater the average change rate of comprehensive evaluation results.

To address the issue of ice and snow accumulation on the winter road surface efficiently， a solution involving the application of an antifreeze sand fog seal layer on the road surface was proposed. Initially， the composition， anti-icing mechanisms， and preparation procedures for the antifreeze sand fog seal layer were delineated. Subsequently， tests were conducted for ice-melting， de-icing， and freeze-thaw cycles to evaluate its ice-suppression and snow-melting efficacy. Concurrently， tests for permeability resistance， skid resistance， and abrasion resistance were performed to assess its road performance. Findings suggested that the ice prevention of antifreeze sand fog seal layer was accomplished via the establishment of a hydrophobic surface and the controlled release of de-icing agents， with the optimal proportions of hydrophobic and anti-icing agents being 8% and 40%， respectively. The material not only significantly reduced the accumulation of ice and snow on the road surface but also sustained its de-icing efficacy following freeze-thaw cycles， thus demonstrating both effective ice-suppression and durability. Additionally， the seal exhibited good permeability and abrasion resistance； albeit a slight diminution in the skid resistance of the original road surface was noted， the seal still conformed to the requisite technical specifications.

Because there is vertical pressure on the upper part in the field test of local mortar alkali aggregate， in order to study the mechanical properties of mortar alkali aggregate reaction， a mesoscopic numerical simulation of mortar alkali aggregate reaction was carried out. The sample was prepared by mortar bar rapid method， and the expansion rate， compressive strength and flexural strength of mortar sample were measured. The mortar sample was regarded as a three-phase composite material composed of aggregate， interface transition zone （ITZ） and mortar matrix， and the alkali-aggregate reaction expansion model and alkali-aggregate reaction strength model were established to simulate the compressive and flexural properties of samples with different sizes at different ages， which solves the problem of high compressive and flexural strength of mortar， and obtains the conclusion that the simulation results are basically consistent with the test results. The simulation has good practicability.

A parameter-adaptive fast non-singular terminal sliding mode control method is proposed to address the problem of the three-dimensional overhead crane anti-sway control system， which is susceptible to unknown factors from parameter perturbations and external disturbances. Considering the uncertainties such as the time-varying of payload mass and rope length in the dynamics model of the three-dimensional overhead crane system， as well as external disturbances， this method is based on sliding mode control theory. A composite sliding surface that accounts for both position deviation and payload swing angle is constructed. Moreover， a parameter-adaptive law to automatically tune the controller gains based on the system's errors is designed and ensures that displacement errors and swing angles converge to zero within a finite time. When compared to existing control methods through simulations， the proposed method allows the overhead crane to reach the desired displacement within 3.5 s without overshooting. It effectively suppresses payload swinging， ensuring that the maximum swing angle does not exceed 3 deg. Additionally， it rapidly eliminates any residual swing of the payload. These results indicate that proposed method enhances the dynamic response performance of the overhead crane system， and improving the robustness and disturbance rejection capabilities of the control system.

Siamese network based trackers have two branches which are independent of each other and lack of infor-mation interaction. So it cannot accurately and robust tracking under the challenges of target occlusion and similar object. To solve this problem， an anchor-free target tracking algorithm based on multiple attention mechanism was proposed. Multiple attention mechanism was used to encode the target template and search area features. After improving the feature significance through self-attention mechanism， mutual attention mechanism was used to aggregate the feature interaction between target template and search area， which strengthens this algorithm's discri-mination ability between target and background. At the same time， the anchor-free mechanism was used to complete the end-to-end visual target tracking task pixel by pixel， avoiding the disadvantages of human intervention caused by the anchor frame mechanism. Extensive experiments are conducted on many challenging benchmarks like OTB50， OTB100 and GOT-10K. These results show the anchor-free target tracking algorithm based on multiple attention mechanism proposed has strong robustness against the challenges of target occlusion and similar object， and effectively improves the precision rate and success rate of the tracking algorithm.

The proposed method selects appropriate features by filtering punctuation marks， constructs vectors， segments two or more words to form specific semantics， and labels parts of speech to identify relative parts of speech； Utilizing the Latent dirichlet allocation （LDA） model to establish correlation between topics and documents， clarify document topics， and reduce data feature sparsity； The Bidirectional long short-term memory-conditional random field （BR-BiLSTM-CRF）model detects the boundaries of text named entities through a bidirectional LSTM model， which is combined with the output entity types of the chain conditional random field layer. After adding features of vocabulary and parts of speech， the overall sequence entity edge of the text is detected. The network parameters are corrected using cross entropy and gradient descent until the error does not exceed the specified value， achieving text named entity recognition. Through experiments， it has been proven that the proposed method has fast recognition speed， high accuracy， and strong overall performance. The proposed method can better recognize language through computers， clarify the part of speech of text， and improve the accuracy and efficiency of named entity recognition.

To fast respond the large scale RS image search task， the deep generative adversarial learning ordinal cross entropy Hashing （GANOCH） was proposed. Firstly， the ordinal relation preserving objective function was established based on cross entropy， then the triplet similarity relationship owned the same probability distribution in different spaces. As a result， the ability of preserving ordinal relation is enhanced. Then， to fix the NP hard problem， the discrete binary encoding was relaxed to the continuous representation. Furthermore， the quantization loss function was designed based on triplet ordinal cross entropy and the loss caused by the continuous mechanism was minimized. Finally， a supervised deep generative adversarial network was established to synthesize RS image dataset according to both label matrix and random noise. Thus， the over fitting problem is avoided. The final ANN search comparative experiments were conducted on three large scale datasets， and the proposed GANOCH achieved the best performance.

In object-level hyperspectral image modeling with limited leaf samples， a multi-scale three dimensional-one dimensional cascaded convolution neural network model was proposed. Firstly， in 3D convolution neural network （3D-CNN）， the dilated convolution was embedded to increase the receptive field of convolution kernel， and a multi-scale 3D-CNN network was constructed to extract and fuse spectral-spatial joint features of different scales， to improve the model performance without increasing the number of network parameters. Then， the one-dimensional convolutional neural network （1D-CNN） was cascaded to the optimal multi-scale 3D-CNN network， to further reduce the computational complexity and overfitting degree of model. Finally， on two datasets of chlorophyll content regression for basil leaf and drought stress recognition for pepper leaf， the optimal network architecture was explored， with comparison to a series of baseline CNN models. Experimental results showed that the proposed model can effectively improve the generalization performance and reduce the computational complexity for both regression and classification tasks of leaf hyperspectral images under the condition of small samples.

To reduce the collision probability between robots and obstacles and improve the efficiency of robots， a deep learning based obstacle avoidance planning algorithm for indoor navigation robots was proposed. Firstly， through the indoor navigation robot navigation system， combined with deep learning， the detection and recognition capabilities of moving and non moving obstacles in the robot's environment were improved， thereby obtaining practical reactive obstacle avoidance navigation information that is more in line with the actual scene. Then， using this information， a simulation map was constructed， and an optimal task execution route was selected within the simulation map， the problem of difficulty in obstacle avoidance planning caused by disorderly and irregular obstacles was solved， and obstacle avoidance planning for indoor navigation robots was achieved. The experimental results show that the proposed method has a more reliable obstacle avoidance path， and the obstacle avoidance planning time does not exceed 1.2 s， effectively improving the obstacle avoidance accuracy and work efficiency of indoor navigation robots.

Addressing the issue of difficulty in achieving high accuracy in achieving high accuracy in pancreas segmentation， a dual input 3D network to realize panc-reatic segmentation was proposed. Firstly， the boundary area was highlighted by increasing the context residual information between input slices， and then the attention mechanism was introduced to suppress useless features and increase the expression of effective features， which finally improve the accuracy of pancreatic segmentation. The algorithm was evaluated on NIH pancreas segmentation data set. The experiment-al results show that the performance of the algorithm proposed in this paper is superior to the comparison algorithm.