A passenger car tire 205/55R16 is taken as the research object， and the relationship between tire force transmissibility and tire cavity resonance noise is studied through experiments. It is proposed that the amplitude of tire force transmissibility has the ability of representing the level of tire resonance cavity noise. A tire force transmissibility simulation model with sound-absorbing material is established using finite element method， and it is found that the mechanical parameters of porous sound-absorbing material have more significant influence on the noise reduction effect than the sound absorption coefficient. The optimal Latin hypercube test method and the Kriging approximate model are selected to explore the influences of the density， elastic modulus， width and thickness of the porous material's on the amplitude of the tire force transmissibility， and the multi-island genetic optimization algorithm is chosen to optimize the above four parameters of the porous material. By changing the single width parameter of the porous material， the accuracy of the simulation results is verified through the tire force transmission test and the falling noise test. The results show that compared with the original tire， the amplitude of tire force transmissibility of the optimized tire is significantly reduced， and the cavity resonance noise of the tire is also effectively improved.

The straight-line driving speed of tracked vehicles driven by dual motors will often lose stability due to the drastic changes of road structure and parameters. At the same time， the different driving resistance on both sides of the vehicle will also lead to different driving motor speeds and vehicle deviation. In view of the above problems， a layered control strategy for vehicle straight-line driving is proposed. The upper layer controls the stability of vehicle straight-line driving speed， and the lower layer controls the speed synchronization of bilateral motors， so as to improve the anti-interference of vehicle straight-line driving speed and reduce the driving offset. Around the problem of external driving resistance disturbance， a Luenberger resistance observer is designed， and its observation value is fed back to the upper integral sliding mode speed controller to improve the anti-interference of vehicle speed； The lower controller adopts cross coupling synchronous control to compensate the output torque of motors on both sides， so as to improve the synchronization of motors on both sides. Finally， RecurDyn + Matlab / Simulink is used for joint simulation to verify the effectiveness of the control strategy proposed in this paper.

In order to reduce the influence of hobbing workbench deformation on machining accuracy， the thermal and mechanical deformation of the workbench were studied.A Subpopulation Adaptive Mind Evolution Algorithm-Back Propagation Neural Network was proposed for thermal-mechanical deformation prediction of hobbing workbench. The initial value， weight and threshold of BP Neural Network were adjusted by SAMEA， which effectively improved the prediction accuracy. Combining K-means clustering and grey correlation analysis （GRA）， the thermal error input variables were reduced from eight to three. Aiming at the table deformation caused by cutting force in gear hobbing， the cutting force was represented by the current of machine tool spindle and used as the input variable of the prediction model. Experimental results show that the average prediction accuracy of the proposed SAMEA-BP model is 95.1%， and comparison with other models verifies the validity and generalization of the model.

In order to set the heat flux density， normal load， and tangential load using the rotational load method， as well as the contact heat transfer between the wheels and rails， convective heat transfer， and thermal radiation effects on the wheel surface， thermo-mechanical coupling simulation was used by secondary development of Abaqus. The temperature change curve model and thermal stress prediction model of the highest point of tread temperature during the braking process were provided. This was based on the dynamic changes and distribution characteristics of tread temperature， stress， and strain under various braking procedures. A quantitative relationship between the maximum temperature and the end temperature of the wheel tread and the parameters of braking conditions was obtained by regression. The findings indicate that： under braking conditions of 25 t， 100 km/h， and 0.7 m/s²， the radial stress reaches 625 MPa， the maximum tangential stress is -713 MPa， the maximum radial and tangential plastic strain is -5.47‰ and -3.37‰， respectively， the maximum thermal strain is 5.178‰， and the thermal stress range is 433 MPa. The maximum temperature of the tread is increased by approximately 90°C， 120°C， and 26°C.

To address accurately calculate the buffer force of anti-backward-tilting device （ABTD） of luffing jib tower crane， considering the influence of tower body elastic vibration and boom rigid body rotation on the dynamics of ABTD， a mathematical model describing the relative movement of buffer is established. Combined with the buffer characteristic model of ABTD， the mathematical relationship between the buffer force of ABTD and dynamic characteristics of luffing jib tower crane is established. Taking a luffing jib tower crane equipped with hydraulic ABTD as the research object， multiple parameters of the buffer force calculation model are identified， and the buffer force calculation model is verified by a series of anti-backward-tilting test. According to the experimental value of buffer force of ABTD， the dynamic load factor of unloading load is calculated， and the relevant requirements of dynamic load factor under sudden unloading condition in Design Rules for Cranes are corrected， which provides a reference for the revision of relevant national standards for the design of luffing jib tower crane.

Due to the limitation of sprocket shape and the requirement of on-site measurement for efficiency， the existing external parameter calibration methods are difficult to apply to the external parameter calibration of the hub face. The paper proposes an external parameter calibration method based on the invariance of quadratic curve. In this method， a cylinder machined with concentric rings is placed in the center hole of the sprocket， and the equation coefficients corresponding to the three circles are calculated by the radius of the large and small circles on the ring and the radius of the hub hole of the sprocket. The external parameters of the hub face can be obtained through the above parameters. According to the shape characteristics of the sprocket tooth profile， a screening model of the highest and lowest points on the tip and root circle regions is proposed， and the diameter is obtained by ellipse fitting using the highest point array and the lowest point array. In the experiment， four sprockets of different sizes are used as the measured objects， and the results obtained by the measurement method in this paper are compared with the results obtained by the three-coordinate measuring instrument. The experimental results show that the error of the measurement method in the paper is less than 40 μm.

Aiming at the problems of low vibration energy collection efficiency and poor practical application effect caused by high natural frequency and narrow frequency band of piezoelectric cantilever beam， the characteristics of energy collection and vibration characteristics of cantilever beam were analyzed by the characteristic frequency equation of Euler-Bernoulli beam， and three parameters affecting the natural frequency of piezoelectric cantilever beam were obtained ： effective stiffness， effective mass and free end inertia moment. A spring-connected mass block structure was proposed to optimize the cantilever beam by changing the overall effective stiffness， and the multi-degree-of-freedom spring cantilever beam was obtained by changing the inertia moment of the structure， which reduced the natural frequency of the beam and was verified in the simulation experiment. In the simulation experiment， it was found that the frequency band superposition phenomenon occured？ due to the decrease of the two-order resonant frequency spacing， and the possibility of its generation was explained by the two-order frequency band decomposition. The experimental results showed that the output voltage of the spring-connected mass block cantilever beam structure was increased by 30% and the frequency band was widened by 2.5 times compared with the traditional structure. Although the improvement effect was not obvious compared with the mass cantilever beam， the output voltage can be increased by 60% and the frequency band width can be increased by 3.3 times after the introduction of multi-degree of freedom.

At present， the risk assessment for road network is mainly based on accident data-driven methods. These methods ignore the risk factors differences on different types of roads， such as the ordinary road， tunnel， bridge， and so on. And this will induce the risk assessment for road network is inaccuracy. Aiming at this problem， a traffic accident risk assessment method for road network was presented considering risk heterogeneity. In this method， a road cascading segmentation strategy from coarse to fine was proposed according to the risk factors prior knowledge of different road types and the distribution characteristics of accidents density. This would achieve the accurate division of road segments with heterogeneous risk. Then an adaptive accident risk spatial distribution function was introduced. And a traffic risk assessment model for road segments with heterogeneous risk was established to realize the overall risk assessment of road network. Based on the international open accident database， the proposed method was compared with kernel density estimation with fixed bandwidth. The results show that the presented method has a better ability to discover traffic risk. And the assessment of road network risk is more consistent with actual situation by using the presented method.

A control strategy for left-turn intersection using double-exit line is proposed to solve the ineffective problem of left-turn intersections using single-exit lane when the length of the comprehensive function area is too large. This paper first establishes a capacity calculation model of contraflow left-turn lanes by analyzing the operation mechanism of contraflow left-turn lanes using single and double exits. Then， a signal timing optimization model is constructed which the maximum number of motor vehicles as the optimization objective. The result indicates if the number of conventional left-turn lanes is equal to that of pre-signal exit lanes， the traffic capacity of the two modes is the same. If the number of conventional left-turn lanes is greater than that of pre-signal exit lanes， the capacity of double-exit type is significantly greater than that of single-exit type. Sensitivity analysis shows that the capacity of left-turn intersections increases first and then decreases with the change in the total length of the comprehensive functional area and the proportion of left-turn. The total length of the comprehensive functional area with double-exit lane should be between 80 and 100 m.

To reveal the impact of air-rail passenger perception on overall passenger satisfaction， a three-layer satisfaction index system （the target， intermediate， and indicator layers） was established. The Bayesian network model was constructed by combining GTT and EM algorithms to analyze the influence of nine categories of satisfaction. The results show that the impact of categories of satisfaction on overall satisfaction are personalization， connectivity， operation， reliability， ticketing， comfort， and information service in order of strength to weakness， while accessibility and safety have little effect on overall satisfaction. The result of cross-tabulation analysis quantitatively reflects the complex interrelationship between various indicators， and the subsequent improvement strategies is quantitatively analyzed by the IPA analysis for different indicators， concluding that the current operation frequency， baggage check-in， opening routes， and operation hours need to be improved， while the problems of transfer instructions， seat comfort， convenient ticket collection， fares， transfer quality， and information planning are over-improved.

In order to strictly control the grading accuracy of the continuous mixing equipment supply system， an online aggregate grading detection method based on image processing has been developed. By setting up an aggregate particle collection platform at the belt of the cold storage bin， part of the aggregate is sampled during the fall of the aggregate and the gradation situation at this time is analyzed. A watershed + Harris method is used to divide it and combine the results of 200 pictures to obtain the final synthetic gradation of the asphalt mixture. When the proportion of aggregate sampling is basically the same， the optimal sampling distance that the dividing plate should be adjusted to under different belt speeds is obtained by using EDEM software simulation. By comparing the results of the screening test and the image detection method，the correlation coefficient between the image method and the screening method detection result reached 0.998，which has a good detection effect.

Aiming at the durability of the slurry-soil interface in arid environment under the synergistic effect has become an urgent need for its promotion and application in the field of crack repairs in earthen ruins. Important support sought. The undisturbed site soil was used to carry out indoor simulated fissure grouting with or without pile-mortar synergistic mode， and conducts dry-wet， freeze-thaw， salt-tolerant and alkali-tolerant 4 types of the grout-soil interface formed in different modes for durability test in arid environment. Wave speed measurement and interface direct shear test are carried out on samples at different ages， and the slurry-soil interface bonding is revealed by comparing and analyzing its wave speed， interface mechanical parameters， interface failure mode and shear stress-displacement relationship. The characteristics and laws of performance degradation in arid environment. Results show that the failure form of the interface during the entire age is mainly the bonding failure of the flat interface and the combined failure of the bonding and incisor failure of the undulating interface; under any durability condition， the grout-soil under the four grouting modes （the interface wave speed， surface residual soil thickness， peak shear stress， and interface shear strength parameters） all show a downward trend with the increase of age， and the size relationship is all expressed as grouting with lime piles and grooves>grouting with grooves> grouting with lime piles>grouting alone; the above rules all show that the durability of the dry environment durability of the pile-mortar synergistic bond between the grout-soil interface of the soil ruins is significantly better than the single grouting mode.

In order to study the variation law of the transverse distribution coefficient of the assembled T beam bridge after any girder damage， based on the rigid-jointed beam method， a transverse mechanical model is established， which is suitable for the different positions and the numbers of girder damage. And the change degree of the transverse distribution coefficient is analyzed with the model. The results show that a transverse unloading occurs after the girder damage. And the larger the girder damage degree is or the more the damaged girders are，the larger the reduction rate of the transverse distribution coefficient of the damaged girder is， while the larger the increasing rate of the transverse distribution coefficient of the intact girder is. And the increasing rate is inversely proportional to the cross range to the damaged girder. In addition， when it is rigid connection between the girders， the transverse unloading capacity and the influence area of the damaged girder are larger than hinged connection between the girders. If the damaged girder is not reinforced in time， the intact girder will be damaged in the order of the closer to the damaged one， the earlier the damage occurs.

In order to ensure the safety of the building slab-column structure， the slab-column structure specimen plate was made with steel bars， cement， etc. as raw materials， and the punching failure test was carried out. According to the calculation formula of the joint bearing capacity， the joint bearing capacity under different plate thicknesses， concrete strength， longitudinal reinforcement ratio， and crush ratio was calculated， and the influence of each factor change on the joint bearing capacity was analyzed. The experimental results show that when the longitudinal load of the specimen plate increases to the maximum bearing capacity， instantaneous brittle failure occurs， and warpage occurs at all displacement measuring points. The smaller the value， the greater the punching shear bearing capacity； greatly increasing the thickness of the specimen plate， the reinforcement ratio of reinforcement， the concrete strength， and reducing the punching collapse ratio can improve the punching shear bearing capacity of the specimen plate. When the reinforcement ratio increases from 0.6% to 1.4%， the maximum punching shear bearing capacity of the node can be increased by 29.2%. When the punching ratio is 7 and the reinforcement ratio is 0.6%， the increase of the concrete strength is small， and the increase in the punching shear bearing capacity of the node is small. To a certain extent， the bearing capacity of the building slab-column structure is improved.

A three-dimensional meso-scale numerical model of corroded reinforced concrete（RC） beams was established by random aggregate model in the present study， and the connector element was used to capture the changes of interaction at the interface during loading. The rationality of the model was verified by experimental data. The interfacial bond-slip distribution and deformation coordination coefficient of corroded RC beams during the loading process were analyzed， and the effect of bond degradation on bearing capacity was discussed. Results show that the connector element can describe the bond stress and relative slip at the interface during the loading process， which is an effective method for simulating the interface behavior between corroded steel and concrete. The maximum bond stress at the interface decreases with increasing corrosion loss， a 30% of corrosion loss leads to a decrease of 71% in the maximum bond stress at the ultimate load. Corrosion can accelerate the relative slip at the interface of corroded RC beams， and the relative slip at the interface moves from the loading point position to the shear-flexure section. The ultimate deformation coordination coefficient at the interface decreases with the increase of corrosion loss. When the corrosion loss reaches 30%， the ultimate deformation coordination coefficient is 0.479， and the bearing capacity decreases by 55%.

Aiming at the problem that the existing access control methods do not consider the features of cloud computing privacy protection， lack dynamic privacy authorization mechanism， and thus can not protect privacy data adaptively in the running process， an access control model TBRBAC for SaaS layer privacy protection in cloud computing is proposed by extending RBAC model with the trust degree and privacy use behavior. On this basis， a trust degree evaluation and dynamic update mechanism for SaaS services is proposed. Then， the architecture of adaptive privacy access control system based on TBRBAC model， its execution process and authorization analysis algorithm are given. The rationality of the adaptive access control process is also discussed. Finally， the feasibility and effectiveness of this method are illustrated by example analysis and experimental verification. This method can achieve dynamic privacy authorization and fine-grained privacy access control at run-time， and enhance the security protection of private data in cloud computing environment.

In the process of parallel density clustering， the boundary points of clusters with different densities are divided fuzzy and there is data noise， which affects the clustering performance and makes the clustering results subject to the influence of local optimization. Therefore， a parallel density clustering algorithm MCS-KDBSCAN （maprule based parallel maximization cuckoo search K-means DBSCAN） based on MapReduce and optimized cuckoo algorithm is proposed. Firstly， the algorithm combines the strategy KDBSCAN（K-means DBSCAN）， which is based on the idea of nearest neighbor and inverse nearest neighbor in k-means. By calculating the influence space of each data point， the expansion conditions of clustering clusters in DBSCAN algorithm are redefined to avoid the problem of fuzzy boundary points of clustering clusters with different densities； Then， combined with the nearest neighbor idea in KDBSCAN density clustering， a feasible iterative noise point processing strategy is proposed to reduce the impact of noise points in data on the performance of clustering algorithm； Secondly， the optimization and improvement strategy MCS （maximization cuckoo search） based on the traditional cuckoo algorithm is proposed. By attenuating the weight of the probability of finding nests， with the increase of the number of iterative searches， the convergence speed of the algorithm is improved， and the influence of local optimization on the clustering results is solved； Finally， combined with MapReduce， a parallel density clustering strategy MCS-KDBSCAN is proposed. By parallelizing the operation of density clustering algorithm， the communication burden of local optimal solution transmission of parallel clustering algorithm is reduced and the performance of the algorithm is improved. Experiments show that the proposed mcs-kdbscan parallel density clustering algorithm is superior in clustering accuracy and clustering running time.

Due to the lack of dimension reduction in the process of cube abnormal data detection， the detection accuracy of abnormal data in the cube is low， the error detection rate is high， and the detection time is long. Therefore， a cube abnormal data detection method based on spectral clustering is proposed. Cluster the data in the multidimensional data set through Laplace matrix， preliminarily classify the data， use LLE algorithm to reduce the dimension of the classified data， express the high-dimensional data set with eigenvectors， remove the redundant information in the multidimensional data set， input the processed multidimensional data set into the support vector machine model， and complete the detection of abnormal data according to the calculation of regression estimates. Experimental results show that the proposed algorithm has higher accuracy， lower false detection rate and shorter detection time.

The current deep learning-based compressed sensing and reconstruction network mainly has the following problems： the single-channel reconstruction network does not deeply explore the multi-scale features of the image， lacks a feedback mechanism for the reconstruction network， and the reconstruction network lacks correlation with the measurement matrix， which restricts the further improvement of the reconstruction quality. Therefore， an image compressed sensing and reconstruction algorithm based on multi-scale generation adversarial network is proposed， which extracts the multi-scale information of images through multi-channel residual blocks， joins the discriminant network to form a feedback to the multi-scale generative network， and then trains the full convolutional measurement network jointly with the reconstruction network to improve the image reconstruction quality. Experimental results show that the reconstruction accuracy of the proposed method is improved by 2.02-4.09 dB compared with ISTA-Net+ method under three sampling rates.

In order to improve the efficiency of the multi-object tracking method， a joint detection-apparence network based on anchor aligned convolutional feature， called AAC-JDAN， was proposed. On the basis of the object detection network YOLOv3， an anchor transformation network and an anchor aligned convolutional operation was introduced， so that the network can detect the rotated objects， while alleviating the problem of the weak correlation between the apparence feature extracted by the exsiting joint network and the rotated objects； by adding an apparence feature extraction branch in the detection network， two subtasks of object detection and object apparence feature extraction were combined in a multi-task joint learning manner to realize the sharing of the low-level feature， and the apparence feature vectors can be extracted along with the corresponding detected objects， which improves the overall efficiency of the tracking algorithm. A fast online data association method was proposed to realize the efficient tracking of multiple rotated objects in the video. The similarity matrix between the incoming detections and the trajectories was calculated with the object apparence feature extracted by AAC-JDAN and the motion prediction result given by the Kalman filter， and the matching was done by the KM algorithm. When tested on two public datasets and a custom dataset， the TPR， MOTA， and IDF-1 reached 80.4%， 71.3%， and 69.5%， respectively， and the framerate reached 20 frames per second， this showed that the proposed method achieves a better balance in the speed and accuracy of tracking.

To overcome the shortcomings like computational expensive and inability to generate structures with explicit geometry of traditional topology optimization algorithms represented by the Solid Isotropic Material with Penalization（SIMP） method and the level-set method， a deep-learning-based two-stage approach for real-time explicit topology optimization was proposed， which combined deep learning with the moving morphable components（MMC） method. In the first stage， a deep learning model was used to replace most of the time-consuming finite element analysis. The second stage performed a small number of iterative fine-tuning of the structure predicted by the deep learning model to generate the final optimized structure with explicit geometry. A relatively general data set was used to verify the feasibility and effectiveness of the framework quantitatively and qualitatively， and the relationship between the training degree of the deep learning model in the first stage and the quality of the structure generated as well as the total time consumed was studied. Experimental results show that this approach can save more than 90% of the computing time while maintaining the quality of the topology optimization structures generated.

Prevalent scenario of support vector machine （SVM） in data classification exists the low recognition accuracy problems， this paper proposes a feature selection model for synchronous optimization of SVM with improved Whale optimization algorithm. Firstly， the Levy flight strategy was used to perturb the spiral update position of the whale optimization algorithm and the reflection operation of the simplex method to improve the reflection point solution of the elite individuals in the population. With extensive empirical validations on standard function， the proposed method can effectively improve the convergence speed and calculation accuracy. Secondly， SVM kernel parameters and feature selection targets were the co-optimization objects effect for obtaining the optimal kernel parameters and the corresponding optimal feature subset. Finally， feature selection simulation experiments were conducted on UCI standard data set and real breast cancer data set， and the average classification accuracy， average fitness value， fitness standard deviation and the number of selected features were evaluated and analyzed. Compared with traditional support vector machines， the classification accuracy of real breast cancer data set is improved by 11.053%.

Aiming at the problem of insufficient acquisition and use of the global relationship between items in various current session based recommendation models， a session based graph neural network recommendation model combined with global potential information is proposed. The model constructs session graph and global graph according to all session sequences， introduces the spacing information between nodes in the sequence and the contribution of adjacent nodes of sequence nodes to each other， obtains the final session representation through model training， and predicts the next interaction behavior. The experimental results show that fully mining the global potential information in the recommendation algorithm combined with graph neural network can effectively improve the accuracy of the recommendation algorithm. This improvement has certain guiding significance for improving the performance of session based graph neural network model.

In order to accurately analyze the problem of soil temperature characteristics， this paper proposes a prediction model based on attention mechanism of multi-channel long and short-term memory network fused with ARIMA algorithm. The attention-based multi-channel long- and short-term memory network is used to extract important temporal features at different moments of the long and short term， and the ARIMA time series model is used to extract linear features to take advantage of predicting soil temperature more accurately. To validate the proposed model， four evaluation metrics， root mean square error， mean absolute error， mean square error and coefficient of determination， were tested at two weather stations in Switzerland （Laegern and Fluehli weather stations） for the next 6， 12 and 24 hours， while soil depths on 5， 10 and 15 cm soil temperatures. Compared with thse autoregressive integrated moving average model， the long and short-term memory network and the fully connected network， the proposed model has the optimal performance， especially the most significant improvement in the soil temperature model for Fluehli station （10 cm soil depth） during the next 6 hours； the highest relative coefficient of determination value of 0.9965， the lowest root mean square error of 0.3414， the average absolute error of 0.2310 and mean squared error of 0.1165.Therefore， the proposed model can be used as an alternative alternative method for soil temperature estimation.

In this paper， aiming at the pseudo-random transmitting current of 200 kHz clock frequency， comprehensively considering factors such as sensitivity， signal-to-noise ratio， the resonant frequency of the air-core coil sensor is increased to above 150 kHz when the noise level is as low as 4.86 nT/s； The receiver is based on the PXIe bus， which can realize the acquisition of three-component electromagnetic data， transmitting current and various auxiliary information. The sampling rate can reach up to 1.25 MHz， and the dynamic range is 112.6 dB. Through the integration of multi-channel signals， improved multi-information single-file storage method and data storage protocols through composite time stamping， reliable multi-information storage under massive data is realized. Compared with the CHTEM-Ⅱ airborne electromagnetic recording system， the throughput is up to 6 times. The test results of various performance indicators prove that the receiving system has a high degree of completion.

In view of the arbitrariness of the arrival signal， adopted the basis function sampling to establish the SFC model of MIMO system. Combined any spectral distribution of the approximate fitting algorithm was proposed， extended the calculation method of small angle spread to equivalently fitting arbitrary distribution. The measured data of urban propagation channel were verified by the piecewise exponential nonlinear fitting analysis. Therefore， the measured channel of MIMO under the approximate fitting solution was carried out， the impact of the sampling number and the weighting coefficient of the basis function on the accuracy was discussed. The research results show that the fitting method has a high degree of fit for the integration of the spectral distribution of the arbitrary arrival signal and the measured channel data. The proposed approximation algorithm can accurately fit the fading-related channel characteristics of the MIMO multi-antenna system， and can greatly reduce the complexity of calculation， improve the computational efficiency and save the simulation time. This method can effectively study the channel characteristics of MIMO systems in special scenarios， and can be extended to the analysis of Massive MIMO in multi-dimensional spatial domains.

Signal processing in the background of α noise is a hot issue in this field， but it is very difficult to directly obtain the relevant information of characteristic exponent of α noise in actual working conditions， which makes the application of fractional low-order statistical algorithm become particularly difficult. An estimation method based on the plus property of α-stable distribution and geometric power was proposed regarding the issue above. Firstly， the plus property is used to determine the relationship between several independent variables with the same α-stable distribution and the distribution of their sum. Then the characteristic exponent is estimated by using the characteristics the geometric power between the original variables and their sum-distribution variable. The experimental results show that this algorithm does not need to obtain the range of the characteristic exponent in advance. And it can be accurately estimated in the range of 0—2， the maximum root-mean-square error of the estimation result is only about 0.1 and the deviation is only 0.02 when it is estimated the sea clutter data， which can provide a priori information under the signal processing problem based on α noise.

In order to meet the requirements of the Internet of everything and physical layer security， a design method of secure beamforming assisted by intelligent reflecting surface is proposed to realize active defense against the communication security of status update system with multiple eavesdroppers and multiple legitimate users.Meanwhile， the requirement of information freshness in the status update scenario is considered， so the information age performance metrics of average confidentiality age and interruption probability of confidentiality age are introduced. Under the constraints of transmit power， state update probability and IRS reflection phase shift， the beamforming matrix， IRS phase shift and state update probability are jointly optimized to ensure state update while maximizing system security. For the dynamics and complexity of the system and the challenge of non-convex optimization， a secure beamforming method based on double deep Q network algorithm is proposed to realize the optimal beamforming strategy for eavesdroppers in dynamic environment. The simulation results show that the proposed secure beamforming method based on double deep Q network can significantly improve the security of IRS assisted state updating system.

In order to meet the demand of precise and quantitative fertilization in modern agriculture， this paper obtains the reason why the screw fertilizer ejector can not realize uniform fertilizer discharge through theoretical analysis. a new type of fertilizer discharge device based on the principle of screw extrusion conveying with full filling of fertilizer is designed， and the parameters affecting the uniformity of fertilizer discharge were obtained through theoretical analysis. The installation inclination and length of fertilizer outlet were taken as the test factors， and the variation coefficient of fertilizer uniformity and the average speed of fertilizer particles were taken as the test indexes. The quadratic general rotation combination test with two factors and five levels is designed， the results showed that the length of fertilizer outlet and the installation angle of fertilizer outlet had a very significant effect on the variation coefficient of fertilizer uniformity （P<0.01） and （0.01<P<0.05）， respectively. The installation angle of fertilizer outlet and the length of fertilizer outlet had a very significant effect on the average velocity of fertilizer particles （P<0.01）， when the installation inclination of the fertilizer outlet is 40° and the length of the fertilizer outlet is 105 mm. The variation coefficient of fertilizer uniformity ≤10% and the average speed of fertilizer particles ≥15.5 mm/s can be obtained. The relative errors of bench test and simulation test are 2.75% and 1.75% respectively， the verification results are consistent with the simulation results. Under different rotating speeds， the uniformity variation coefficient of extrusion screw fertilizer discharge is 32.91% lower than that of single screw fertilizer discharge， and the amount of fertilizer discharged in a single cycle at different rotating speeds is 6.70% lower， the optimized fertilizer ejector has good fertilizer uniformity and small fertilizer loss. This study can provide reference for the optimal design of fertilizer distribution apparatus with screw.

Aiming at the problems of complex assembly process， high cost and long product development cycle of tractor gearbox， this paper studies the assembly technology of tractor gearbox based on virtual reality， and develops a virtual assembly system of tractor gearbox with characteristics of immersion， imagination and interaction. Firstly， the construction method of gearbox virtual reality model was studied， and the assembly information of tractor gearbox was extracted from CAD system by means of model data conversion， and the assembly information hierarchical model of tractor gearbox was established in the virtual assembly environment. The lightweight virtual reality model was obtained by using Mesh merging and geometric element deletion. Secondly， virtual assembly software is developed in Unity3D environment based on C# language to realize the functions of transmission parts in virtual assembly environment， such as movement， rotation， positioning assembly， scene roaming， sound playback， parts name display， assembly path planning， etc. Finally， Oculus Quest2 head-mounted display was used as interactive equipment to carry out tractor gearbox assembly test. The experimental results show that the developed gearbox virtual equipment system runs smoothly， the model interaction is accurate and the user operation experience is good， which helps to shorten the gearbox virtual assembly cycle and reduce the development cost.