In the process of long-distance cargo transportation of heavy truck， the main driving condition of the vehicle is the high-speed cruising condition with little change in speed range. In order to further improve the fuel saving rate of the vehicle under this condition， a predictive energy saving algorithm with the objective of minimizing engine fuel consumption is proposed， which takes the P2 configuration single-axle parallel hybrid truck as the research object. Based on the change of road slope in front of the vehicle， considering the influence of speed change and energy distribution on fuel consumption， the vehicle speed and battery SOC are regarded as system state variables， and the global optimal future vehicle speed trajectory and energy distribution rules for the current road are determined by dynamic programming algorithm. It has been verified that the speed planning and energy distribution results obtained by the predictive energy saving algorithm are reasonable， and the fuel saving rates of hybrid truck with predictive energy saving algorithm are 3.19% and 6.26%， respectively compared with hybrid truck using dynamic programming algorithm under cruise control system （CCS） and pure fuel truck under predictive cruise control （PCC）.

The design of the docking and locking system scheme adopted the concept of "profile sliding guidance + multi-point locking". Static and dynamic analyses of the docking process were conducted using the capture domain theory and dynamic simulations. A scaled-down experimental platform was established for multi-objective docking and locking tests， thus confirming the reliability， stability， and speed of the modular flying vehicle. By controlling the communication system and mechanical guidance， the modular flying vehicle can achieve pose correction under conditions of horizontal deviation within a range of \pm \mathrm{10} mm and heading deviation within a range of ±5 degrees. Experimental results indicate that this docking system ensures the accuracy， stability， and swiftness of module transitions during modal switching.

To solve the problem of vehicle deviate from the planned trajectory caused by the changing radius of curvature of the route， a robust control approach based on H_∞ loop shaping is proposed to improve the automated trajectory tracking performance. Firstly， the vehicle dynamics model is analyzed and processed to provide the generalized interference control system form， and then the function operation of transfer is used to indirectly obtain the required output variables. Secondly， the nominal plant is selected and the nominal condition is given， the system singular values of the plants can obtain desired shapes based on the robust controller and thus can meet the system performance index and robust stability requirements. Finally， the automated tracking ability of the vehicle on the variable curvature paths is tested. The results show that the vehicle has a stable anti-interference ability under the action of fixed order robust controller， when parameters change within a reasonable range， the lateral error of the vehicle can be controlled within 0.2 m， the goal of good automated tracking performance in multi-curvature routes of the vehicle has realized. Therefore， this control method has a widely potential applications prospect and is suitable for the development of intelligent vehicle path tracking lateral control and autonomous vehicle lane keeping control.

To further improve the prediction accuracy of the thermal error model of the feed axis of the gear grinding machine， a thermal expansion modeling method of the feed axis based on principal component regression is proposed in this paper. The slope parameter of thermal expansion is obtained by decoupling the positioning error of the feed axis through a linear fitting， which eliminates the position correlation between the thermal expansion error and the position of feed axis. The regression model between the thermal expansion slope and all the temperature points is established using the principal component regression algorithm. Different from the traditional methods， the principal component regression model does not need additional screening of temperature sensitive points， and the mean value and standard deviation of the root means square errors of the prediction results can reach 2.0 μm/m、0.9 μm/m， which has higher accuracy and stability than conventional methods.

In order to improve the efficiency of hyperparameter setting and its adaptability to the model， and break the high cost and low efficiency of manual parameter setting， a fault diagnosis method for rolling bearings based on the honey badger algorithm （HBA） optimizing bi-directional long short-term memory （BiLSTM） with attention mechanism （HBA-A-BiLSTM） is proposed. Firstly， search the optimal hyperparameter combination of the A-BiLSTM model through HBA. Secondly， the fault diagnosis performance is tested based on the A-BiLSTM model under the optimal hyperparameters. Finally， the generalization ability of the model is tested based on the datasets under different working conditions. The CWRU dataset is used to verify the fault diagnosis effect of the proposed method， which is used the diagnostic accuracy and the confusion matrix to evaluate. It is shown that， compared with other swarm intelligence optimization algorithms， the HBA has better global searching performance and faster convergence speed. The fault diagnosis accuracy of the optimized model has reached 99.5%， which has a good effect， also under different working conditions， it can achieve stable and accurate fault diagnosis performance， and has strong generalization ability.

To address the problem of automatically roughening the sole with some contour error， an elastic roughing method is proposed. Firstly， the sole contour is divided into theoretical contour and manufacturing error. The mathematical model describing the sole contour is established. The theoretical contour of the sole is tracked by the two degree of freedom motion of the sole. The manufacturing error of the sole contour is adapted by the deformation of the grinding head under the contact action. Secondly， combining the force control of the grinding head with the motion control of the sole， the dynamic model of the elastic roughing mechanism is established. The contact force changes between the sole and the grinding head under different grinding head postures are simulated， and the key parameters of the elastic roughing are studied. Finally， the test system of sole elastic grinding mechanism is developed， and the feasibility and effectiveness of the elastic roughing of the sole are verified by experiments.

Aiming at the problem that the slip line defects are easy to occur at the sharp edges in the drawing process of automobile hood outer panel， the process parameters that may cause appearance defects were studied. Firstly， the influence of blank holder force， friction coefficient， drawbead strength and stamping direction on the forming of engine hood outer panel is studied by using the single factor variable method， and the primary and secondary order of each factor affecting the slip line and the approximate range of good drawing forming are determined. Then， orthogonal experiment method and range analysis are used to find the optimal level combination. The research results show that the application of numerical simulation technology can improve the forming quality of engine hood outer plate， reduce the number of die tests， shorten the production cycle and reduce the production cost. The reasonable combination of process parameters can optimize the slip line defects and increase the control method of slip line defects.

Based on the finite element software ABAQUS， finite element models of the friction process of a car front axle rocker bushing with flexible surfaces similar to the head surface of the dung beetle are established and their friction characteristics are verified. By selecting an appropriate hyperelastic constitutive model and changing the size of the convex hulls on the surface of the bushing， the variation law of the radial and torsional mechanical properties of the flexible friction pair are analyzed. It turns out that the design of bionic convex hulls changes the maximum stress and strain positions of the bushing under stress， then it can effectively prolong the service life of the rubber bushing surface； the contact friction force and friction coefficients of the bionic bushings surfaces are obviously reduced， the friction reduction and wear resistance of the type2 model surface is the best.

In response to the traffic congestion caused by a large number of lane-changing behaviors in bottleneck sections of expressways， taking the weaving section of expressway as an example， which is a frequency variable area of traffic flow， this paper analyzes the characteristics of vehicle infrastructure cooperation operation， establishes a dynamic assignment model of traffic flow in the bottleneck section， uses C-V2X， edge computing and other technologies， designs a dynamic allocation system of vehicle flow based on vehicle infrastructure cooperation， and studies its operation mechanism. The traffic flow allocation system is simulated and analyzed using MATLAB cellular transmission model. The simulation results show that the optimization effect of the dynamic allocation system is obvious. The traffic time is reduced by about 11.4% in synchronous flow state and 39.7% in congested flow state. The research results are of great significance to reduce the operation delay in the weaving bottleneck area of the expressway and improve the traffic flow efficiency and expressway capacity.

Regarding the accurate prediction of passenger flow at the entrance and exit of rail transit stations， Considering the land use around the station， Rail transit connection conditions， station attributes and attraction， a two-level passenger flow prediction model is constructed in this paper， station attributes and attraction， a two-level passenger flow prediction model is constructed in this paper， including the passenger flow prediction at the traffic district level based on multiple nonlinear regression and the passenger flow prediction at the entrance and exit level based on the CRITIC method. On the basis of the available data， the total passenger flow of all entrances and exits in the traffic district is predicted， and then be allocated to each entrance and exit. Different types of rail transit stations are randomly selected to verify the effectiveness of the model. The results show that the error between the predicted value of the model and the actual value of the daily entry volume is within 30%， and the average error is 20%， which has a high prediction accuracy.

In order to explore the motion and safety characteristics of the U-turn vehicles at long-range U-turn intersections， based on the measured data of the U-turn vehicles at typical locations， we analyzed the characteristics of the U-turn vehicle's trajectories and traffic parameters with the help of probability theory and statistical knowledge， and established the U-turn trajectory model. We used traffic conflict technology （TCT） and take post encroachment time （PET） as the indicator to study the feature of vehicles security of U-turn vehicles. Three conclusions are drawn from this study. First， there is a significant difference in the headway between long-range U-turn vehicles and those at intersections without opening： the stability of the headway of long-range U-turn vehicles is weak， and the release of vehicles is seriously disturbed by the turning flow； Secondly， when turning around at a long interval U-turn intersection， the starting position of vehicle turning is dispersed and the distribution of U-turn trajectories is quadratic distribution. Third， the setting of long interval turning lanes will increase the interference between traffic flows at intersections， thus increasing the number of conflicts with vehicles in the opposite direction. Cross-conflict accounts for the majority of conflicts in long interval U-turn intersections， and rear-end collision is the main conflict type in U-turn intersections without opening. The above research conclusions are helpful to improve the safety of long interval U-turn intersections.

When the traditional deep learning-based models predict link speeds for the entire urban traffic network， they do not consider the proactive feature （signal control information） of traffic flow， and therefore achieve low prediction accuracy. In order to tackle this issue， this paper proposed a link speed prediction framework， based on generative adversarial network and graph neural network. By adopting a proactive and a reactive prediction module， the generator of this framework is able to encode traffic flow and signal control information at the entire network level. The discriminator is then used to increase the generalizability of the prediction outcome. By comparing its performance with traditional time-series and deep learning-based models in real-world traffic circumstances， it is found that the proposed framework achieved less prediction error （3%-5% RMSE drop） than the SOTA model （ASTGCN）.

Aiming at the problem of large energy consumption caused by traffic infrastructure equipment， traction motor， brake device， etc.， an energy-saving optimization control method based on Sequential quadratic programming （SQP） algorithm for urban intelligent transportation operation was proposed. Based on the operation process of urban intelligent transportation， kinematics analysis of traffic train is carried out. According to the total energy consumption of traffic operation traction， a traffic operation energy-saving control model is constructed， and SQP algorithm is used to solve the model， so as to realize online energy-saving optimization control of urban intelligent traffic operation. The experimental results show that the method carries out energy saving optimization control performance test and regeneration energy utilization test， which verifies the practicability and robustness of the method.

To ensure the safety of curved bridges when the customized transport vehicles （CTVs） pass， a safety assessment method for resistant overturning stability was proposed based on forward and inverse reliability analysis. Firstly， considering two limit states which corresponding to conditions that the bridge cannot satisfy the safety redundancy and overturning abrupt， two levels of safety assessment functions were constructed， and the corresponding target reliability index analysis methods of two levels were proposed. Secondly， based on the inverse reliability theory， a two-level resistant overturning stability safety factor calculation model was established. Finally， the stability and safety of the typical three-span continuous curved bridge against overturning was evaluated. The results indicate that the proposed calculation model is reasonable and effective. When a typical three-span continuous curved bridge is equipped with a single support， it meets the safety requirements for anti-overturning， but has low safety reserves. It is recommended to optimize the design of the support form or take anti overturning measures.

In the mixed transportation network with random charging demand， a bi-level programming model is built to determine the reasonable number of charging piles. The upper model aims to minimize the sum of the net present value of the construction cost， the cost of customer loss， and the cost of exceeding the endurance time； the lower level model aims to minimize the user's generalized travel cost to describe the interaction between the road network flow allocation and the location of charging facilities. The model uses differential evolution algorithm and Frank-Wolfe algorithm nested to carry out traffic equilibrium assignment and charging pile configuration optimization simulation. The results show that the model can effectively simulate and predict network equilibrium traffic flow， random charging flow， queuing loss rate and optimal number of charging piles.

In order to improve the efficiency of horizontal transportation of automated container terminals and reduce conflicts between automated guided vehicles （AGV） in the horizontal transportation， a multi-AGV horizontal transportation scheduling and path planning model with the goal of minimizing makespan is established， a dynamic allocation strategy of road network resources is proposed， and a two-layer algorithm composed of Cultural-Genetic Algorithm and Dijkstra Algorithm based on time window is designed to solve the model. The upper layer algorithm is the Cultural-Genetic Algorithm to optimize the scheduling of AGV， and the lower algorithm is the Dijkstra Algorithm based on time window for conflict-free route planning， which effectively reduces the makespan of horizontal transportation at the automated terminal and reduces the possibility of conflict. By comparing the dynamic allocation policy control method， speed control method and task priority control method based on road network resources， the effectiveness of the proposed method in solving the problem of horizontal transportation scheduling and path planning of automated container terminals is verified.

In this paper， the standard krill algorithm （KH） has the disadvantages of slow convergence speed， insufficient calculation accuracy and easy to fall into local optimal solution for complex problems， an improved krill algorithm （SDEKH） which combines improved differential evolution operator and S-type adaptive inertia weight is proposed in this paper. Through a variety of standard test functions to compare and test a variety of intelligent algorithms such as SDEKH and KH， the excellent performance of SDEKH is verified， and SDEKH is used to optimize the truss structure， and the optimization results of SDEKH are compared with other methods to verify that the optimization efficiency and accuracy are improved， which provides a more efficient and accurate method for engineering structure optimization design.

To improve the computational efficiency and performance of the existing algorithms of arrangement optimization of dampers in energy dissipation structures， a hybrid algorithm of differential evolution and artificial electric field was proposed. This algorithm utilizes Tent chaotic mapping initialization to improve the quality and diversity of population distribution， and introduces a mutation cross selection mechanism to maintain the evolution of dominant individuals in the population and avoid approaching local optimum. Through a test function example and an arrangement optimization example of dampers of concrete frame structure， the optimization performance and convergence speed of various optimization algorithms were compared. The research results indicate that the hybrid algorithm of differential evolution and artificial electric field has good global optimization ability， fast convergence speed， and the optimized structure has good seismic damping performance.

In order to improve the rationality of vehicle charging station layout and reduce resource waste， a new energy vehicle charging station location method based on improved particle swarm optimization algorithm is proposed. Predict the future distribution of electric vehicles， and take the user travel characteristics， traffic density， service radius and other factors as the reference basis for location selection； Taking the shortest distance between the demand point and the charging station as the objective function， set the relevant constraints and establish the location model； Explore the implementation process of classical particle swarm optimization algorithm， and obtain particle velocity and position update formula； Aiming at the problem that the method is easy to fall into local optimum， genetic algorithm is used to improve it； The improved algorithm is used to solve the objective function， set the initial parameters and judgment conditions， increase the particle crossover， mutation and other operations， and improve the quality of particle swarm. When the requirements of iteration times are met， the optimal location of the individual is output， that is， the optimal scheme for the location of the charging station. The experimental results show that the location selected by the proposed method can meet the demand of the objective function， balance the charging demand， and avoid resource waste.

In order to solve the problems of slow cache scheduling and poor performance in multi-core systems， a cache scheduling algorithm optimization method for embedded multi-core systems is proposed. By establishing a multi-core system scheduling model， analyze the overall resource load balance of the system； Calculate the synchronization ratio of the system by expanding the scheduling strategy； By using integer linear programming equations， the system achieves task load balancing， minimizes communication overhead， and satisfies on-chip storage limitations， achieving cache scheduling optimization for embedded multi-core systems. The experimental results show that the proposed method has lower communication overhead and a cache scheduling time between 6.38-12.32 ms， indicating better cache scheduling performance.

At present， there are problems with low accuracy in feature extraction of gesture key points in complex background environments. In order to solve the problems existing in traditional methods， a multimodal gesture key point feature extraction algorithm research is proposed in complex environments. Firstly， the gesture image is enhanced by improving the bacterial foraging （BFO） optimization algorithm； Secondly， background removal is performed on gesture images through conditional generation of adversarial networks； Finally， the GIFT method is used to detect the key points of the gesture image， and the multimodal gesture key poiti-scale dual tree complex wavelet transform method and Gabor filtering method. The experimental results show that the proposed algorithm has higher accuracy and better performance in extracting gesture key point features.

To accurately detect abnormal links in wireless sensor networks， a CNN based wireless sensor network link anomaly detection algorithm is proposed. Design network link anomaly detection function using concurrent multithreading technology， and establish a training model for network link anomaly detection using CNN. Input network link data， convolve network link information， extract network link feature vectors， and analyze network link abnormal behavior through down sampling function processing. Use vector mapping to represent the abnormal part vector， and complete classification detection through Softmax function classifier. The experimental results show that the proposed method can effectively improve the accuracy of link anomaly classification and detection， and it takes a short time.

Aiming at the problems of uneven distribution of grayscale values and high noise in images caused by uneven lighting， as well as poor visual communication effect， a visual information compensation based method for enhancing images with uneven lighting is proposed. Calculate the pixel grayscale values of different targets and background regions in the image， construct a grayscale attenuation sequence， construct an attenuation function to weight each background point around the weakest pixel， and use a homomorphic filtering function to denoise based on changes in the component of light in the image. By using the visual information compensation method to calculate the pixel values of each region in the graph， and converting them into component modes， complementary enhancement is implemented based on the total brightness value and local values. The experimental data proves that the proposed method has good enhancement effect， improved image clarity， and good detail characterization effect.

In order to solve the problem that it is difficult to effectively integrate the features of CNN and Transformer network in the image classification task， this paper proposes an image classification framework based on knowledge distillation： Knowledge distillation image classification （KDIC）. In the KDIC framework， a variety of knowledge distillation methods are designed according to the difference of the network structure between CNNs and Transformer： this method effectively integrates the local features of CNNs and the global representation of Transformer into the lightweight student model， and proposes effective loss functions based on different knowledge distillation methods to improve the performance of image classification tasks. The image classification experiment was carried out on three public datasets， CIFAR10， CIFAR100 and UC-Merced. The experimental results show that the KDIC framework has obvious advantages over the current knowledge distillation method， and KDIC still has good performance and good generalization under different teacher and student networks.

The extraction accuracy of intersecting line weld affects the accuracy of industrial welding， which is an important step of weld technology and one of the key issues in industrial research. In order to solve the problems existing in the traditional method， the intersection line weld extraction algorithm based on convolution neural network and bilateral filtering is proposed. Firstly， the weld image of the intersecting line is grayed and enhanced by the contour wave transformation method； Secondly， the weld image is de-noised by bilateral filtering method； Finally， the intersecting line weld seam is extracted through the full convolution neural network. The experimental results show that the proposed method has higher accuracy， faster speed and better overall application effect.

A convolutional network based facial expression recognition method was proposed to solve the problems of large reference number and weak recognition ability in facial expression recognition. The improved residual module was introduced to reduce the parameters and enhanced the attention to the expression area； The channel-space attention mechanism was used to assign the weights of different dimensions and positions to the expression regions extracted from the network， and the subtle feature information of the key points of expression was focused on； The refinement module was used to further extract the depth feature information. In order to obtain higher accuracy， the joint loss function was introduced to increase the out-of-class distance and reduced the in-class distance to improve the accuracy of expression recognition. The experimental results showed that the average recognition rate was 63.91% and 97.98% respectively， and the parameter was 11.34 M. Compared with VGG network and residual network， the model not only improves the recognition rate but also reduces the redundant parameters.

Aiming at the difficulty of object detection caused by the large target size variation， complex background noise and dense targets in airborne wide-area remote sensing images， this paper unifies the target pixel size of the input image by optimizing the segmentation method， and proposes a multi-scale normalized convolutional neural networks model （MNNet）. To enhance the feature correlation between localities， this paper designs a space global connection block （SGC）， which effectively improves the detection accuracy. For the problem that the parameters of the existing NMS algorithm depend on the empirical setting， this paper proposes a self-adaption non-maxima suppression method （DNMS）， which reduces the difficulty of model deployment. The test results on the RSF dataset show that the average precision （AP） of the model in this paper is higher than that of other models by more than 5.0%， and the detection speed reaches 57.7 fps， which can meet the detection task of remote sensing images.

In order to solve the problem of poor tracking effect of multi-target tracking algorithm in the case of occlusion， a multi-target tracking algorithm based on 3D point cloud detection is proposed. The 3D target detector based on point cloud is used to detect the vehicle target and obtain the location information of the 3D target； The target position in the next frame is predicted by tracking the target position in the current frame through a three-dimensional Kalman filter； The intersection ratio of 3D center point space distance and cross-union ratio of bird's eye view is fused as the weight， and the improved Hungarian algorithm is used for data association； Aiming at the problem of label switching before and after occlusion， a trajectory optimization algorithm is proposed. Experiments were conducted on KITTI dataset， and the vehicle tracking accuracy and tracking accuracy reached 84.71% and 86.63% respectively. Under the same threshold， this method is 6.28% and 0.39% higher than AB3DMOT respectively. Experimental results show that this algorithm can effectively improve the performance of 3D multi-target tracking.

When a sudden malfunction occurs in gas insulated switchgear and is not dealt with in a timely manner， it not only leads to the stagnation of the production line， but also causes considerable economic losses to the enterprise. To avoid such situations， this study proposes a device fault upper limit evaluation algorithm based on time series models and deep learning. This algorithm combines the essence of modern data analysis， aiming to improve the accuracy and efficiency of fault detection. This method first applies empirical mode decomposition technology to handle the irregular fluctuation components in time series data. Through this method， the noise and redundant information in the fault signal are effectively removed， making the fault features clearer and easier to identify. Next， we introduced the long short-term memory network in deep learning algorithms. A simultaneous algorithm was constructed by combining LSTM network with EMD technology. This algorithm can simultaneously utilize the temporal and spatial characteristics of data to more accurately identify sensitive components in fault signals. After obtaining equipment fault data， empirical mode decomposition is performed using a simultaneous algorithm to obtain the intrinsic mode function components of the data signal. Then， extract the fault sensitive components from these components and analyze their relationship values and mutual information with the original data signal. Finally， in the process of evaluating the upper limit of equipment failure， a deep learning model is constructed and the refractive and reflection coefficients are determined to achieve accurate evaluation of the upper limit of equipment failure. Experimental tests have shown that the proposed algorithm has achieved ideal results in evaluating the upper limit of equipment failures. The obtained curve has a high degree of fit with the actual result curve， proving that the algorithm can scientifically evaluate the upper limit of equipment faults and has certain application value. This algorithm can not only be applied to the field of fault detection in gas insulated switchgear， but also be extended to other similar industrial equipment， contributing to the safety production and economic benefits of enterprises.

Aiming at the existing radiographic report generation methods that mainly focus on the guiding role of the preceding information in the report generation process and ignore the role of the following information in the report description generation， which has the problems of incomplete semantic information and missing key information， anterior-posterior memory matrix model for chest radiology image report generation（RadRG） is proposed. In order to obtain richer semantic information in the generation process， anterior-posterior memory matrix（MA） is proposed， which uses the memory matrix to record the schema information and semantic information of the report， and uses a gating unit to regulate it in order to prevent the gradient from disappearing or the information explosion in the training process. The model in this paper is tested on the IU X-Ray and MIMIC-CXR public datasets， and the experimental results show that the model in this paper achieves a 2.6% improvement in the BIEU metrics compared with the existing mainstream models.

In view of the problems that rail transit signals are prone to error and distortion， the design of rail transit signal joint control system based on variable structure PID control is proposed. First， the rail transit signal joint control system is designed. The system is divided into traffic command and dispatching module， interlocking module， section blocking and train operation control module， and signal detection and monitoring module. Then， LTE technology is introduced into train-ground communication， To improve the safety and reliability of the system， a variable structure PID fuzzy neural network joint control algorithm for rail transit signal is proposed to realize rail transit signal control. Experimental results show that the proposed method has higher throughput and transmission success rate， and lower average end-to-end delay and bit error rate.

In order to improve energy efficiency and reduce power consumption， a non-orthogonal multiple access （NOMA） power allocation algorithm based on system energy efficiency maximization is proposed for NOMA systems in cellular networks. The multi-user NOMA power distribution problem is transformed into system capacity maximization problem by analyzing the mathematical model under certain conditions. When the conditions are not satisfied， the subgradient method and obstacle function method are used to solve the problem. Simulation results show that the proposed algorithm has significant performance improvement compared with existing methods.

To achieve real-time holographic perception of vehicle trajectories on all sections of highways， this paper aims to explore a cost-effective and easily deployable method for holographic perception of vehicle trajectories. Initially， perception devices composed of laser range sensors， distributed along the roadside， are used to collect discrete trajectory profile information of vehicles in real time. Subsequently， by conditional matching， the information from adjacent detection profiles is matched and associated to obtain the sequence of trajectory points for each vehicle. Finally， cubic spline interpolation is used to reconstruct the complete trajectory of the vehicles. Verification shows that the reconstructed vehicle trajectories are fundamentally consistent with their actual trajectories， demonstrating the feasibility of this method in theory.

This paper presents an experiment investigation of acoustic properties for NACA0012 airfoil at different freestream velocities with 0°and 10°angle of attack（AOA）. The acoustic wind tunnel test results show that at the freestream velocities of 30-60 m/s and AOA= 0°， the dominant noise is broadband noise， but at AOA=10°， the dominant noise is switched into tonal noise， and the tonal frequency is close to the Arbey acoustic feedback model. In order to eliminate the tonal noise of airfoil， a flow control method using a ridge-like structure to break the acoustic feedback loop is proposed， and the influence of the position and arrangement of the structure on the noise reduction performance is discussed. The results show that the closer the structure is to the leading edge of the airfoil， the more obvious the noise reduction effect is. With the increase of speed， the ridge-like structure in the back position no longer has a noise reduction effect in turn. In addition， the generation of tonal noise seems to be only related to the flow state of the pressure surface. When the flow state of the pressure surface is broken by the ridge-like structure， an effective feedback loop cannot be formed. Finally， based on the Arbey acoustic feedback theory， the differences in noise reduction performance of ridge-like structures are explored. It was found that when the ridge-like structure is located before the maximum speed point of the blade pressure surface， it can effectively intervene in the flow field to suppress the generation of monophonic noise.

Except the loose soil on the surface of Mars， there is a kind of special type of soil names Crusty-cloddy， where the ground surface is brittleness and the underlying is loose sand. That is deceptive for Mars Rover. This article take theMars-1and Mars-3 simulant Mars soil as the raw material， to formed hard-shell structures under the nature dry， then studied on the mechanics of these crust， the results showed that the crust of fine particles Mars-1 soil is best when the liquor is saturated NaCl， which has the biggest area， strongly integrity and small thickness； while the Mars-3 crusted the best when liquor is 10% MgSO₄， which has a small area， large thickness but weak integrity. The crusting velocity of Mars-1 is faster than that of Mars-3. Concerning about loading properties， the peak broken force for crust of Mars-1 decreased with the increasing of the concentration of NaCl but increased with the increasing of the concentration of MgSO₄. Two kinds of crust have the biggest penetration resistance under the liquor of 10%NaCl. For Mars-1， the penetration resistance increases with the increasing of the concentration of NaCl but decreases with the increasing of the concentration of MgSO₄.The above research can provide the reference and basis for the planet ramp mechanism optimization design and performance evaluation.