The cooperative platoon control of connected and automated vehicles with model uncertainties and external disturbances is investigated. A fixed?time disturbance observer （DO） is proposed， with which the compound disturbance （i. e. ， model uncertainties and external disturbances） can be estimated accurately within settling time. Based on the DO， backstepping method and fixed?time theory， a novel controller is further designed to ensure that all signals of the closed?loop system have fixed?time stability， which ensures that the tracking error between vehicles converges to zero within a settling time， and the convergence time only depends on the controller parameters. Through Lyapunov stability theory， both individual vehicle stability and string stability are guaranteed. Finally， a simulation of five?vehicle platoon control verifies the effectiveness of the proposed scheme.

Significant progress has been made in pedestrian trajectory prediction， but most of the existing methods are constrained by the limited on?board computing resources. How to achieve efficient pedestrian trajectory prediction in autonomous vehicles is still insufficient. To solve this problem， a lightweight pedestrian trajectory prediction algorithm is proposed， which uses convolutional neural network （CNN） and graph convolutional neural network （GCN） to process and integrate multimodal information. Firstly， a multi?scale feature processing module is designed based on CNN. Multiple convolution modules are used to capture the features of pedestrian tracks and scene information at different time and spatial scales. Then， a feature integration module is constructed based on GCN， which is used to efficiently integrate the spatial?temporal relationship between trajectory and scene features and obtain multiple prediction representations. Finally， multiple prediction representations are integrated to obtain pedestrian trajectory prediction results. Experiments on PIE and JAAD datasets show that the proposed method achieves competitive and optimal prediction performance with the least network parameters， respectively， which verifies the effectiveness of the proposed method. Compared with the previous lightest method， the parameters are optimized by 73%.

Homomorphic encryption （HE） is severely limited in its practical deployment due to low execution efficiency and the inability to ensure data integrity， particularly in scenarios with strict latency requirements. To address such issues and enhance general applicability， a new HE scheme is proposed. To improve execution efficiency， a multithreaded matrix multiplication （MMM） algorithm is designed. With the MMM algorithm， encryption tasks can be decomposed and distributed across multiple threads for parallel execution， thus achieving acceleration. To tackle data tampering in malicious server environments， a verifiable encryption mechanism is designed using zk-SNARK techniques to protect the integrity of ciphertext in outsourced computations. By combining MMM， an efficient verifiable fully homomorphic encryption based on zk-SNARK （zk-VFHE） was developed. Theoretical analysis and experimental results demonstrate that zk-VFHE outperforms similar protocols in terms of both execution speed and security.

A formation obstacle avoidance strategy is proposed based on the improved artificial potential field method to optimize the cooperative obstacle avoidance path of multi?UUV in an unknown underwater environment， which is prone to the local minimum problem and the loss of formation communication due to formation dispersion. Firstly， the formation control between multi?UUV is established based on the pilot?following method， and an improved artificial potential field function in exponential form is used to solve the emergency steering problem caused by the sudden change of the potential field in the traditional form. Meantime， while the effects of forward speed and deviation angle are introduced in the repulsion field， and the deviation degree repulsion combined with the trim angle is introduced to optimize the final obstacle avoidance control command， and a virtual point guidance method is used to solve the local minimum problem. Finally， a cooperative obstacle avoidance strategy for UUV is designed to optimize the formation deviation and recovery time， taking into account the effect of formation. The experimental results verify the effectiveness of the proposed algorithm in multi?UUV formation obstacle avoidance.

During the production and processing of aluminum profiles， multiple types of surface defects such as unclear features and varying scales may generate. In response to the problems of low accuracy， poor real?time performance， and strong subjectivity in existing manual sampling method， an improved surface defects detection algorithm is proposed for aluminum profiles based on PP-PicoDet-XS.The SimAM attention was embedded in the backbone to enhance the ability of extracting deep effective features. The SIoU（Scylla intersection over union） loss function is used to optimize the training process to improve the positioning ability of the prediction boxes. The quantization and distillation were used to compress the model to improve the inference speed. The results show that the improved algorithm achieves a mean average precision of 98.93% at intersection over union（IoU） threshold of 0.5， and 57.60% across IoU thresholds ranging from 0.5 to 0.95， which is 1.73% and 4.13% higher than the uncompressed original model. Deploying this algorithm on the Snapdragon 865 mobile platform for inference， the inference speed can reach 116.82 frames per second， which is 47 frames per second higher than the uncompressed original model.

To address the low solubility and slow dissolution rate of metal oxides at low temperatures， a porous Bi₂O₃/CNTs （carbon nanotubes） composite pellet was used as the cathode， and a graphite rod served as the anode to obtain Bi/CNTs composites through direct electrolysis in a deep eutectic solvent electrolyte composed of choline chloride and ethylene glycol. The electrochemical behavior of the electroreduction process of solid Bi₂O₃ was investigated via cyclic voltammetry. A single pair of redox peaks was observed on the cyclic voltammogram， indicating that the electroreduction of solid Bi₂O₃ to Bi is a one?step reduction reaction. The effect of electrolysis voltage on the electro?deoxidation of Bi₂O₃ was studied through constant voltage electrolysis. XRD and EDS analyses confirmed that Bi₂O₃/CNTs were completely reduced to Bi/CNTs at 2.6 V and 80 ℃ with 4 hours. SEM results revealed that increasing the electrolysis voltage led to larger particle sizes and more porous surface structures in the final products. The study provides a simple and environmentally friendly method for preparing metals by direct electrolysis of metal oxides using deep eutectic solvents.

To effectively utilize by?product coke oven gas and meet the diverse energy demands of users， a distributed multi?energy cogeneration system fueled by coke oven gas was developed. The system incorporates primary energy consumption， environmental compensation costs， and operational and maintenance expenses to create an optimized operational model for efficient energy collaboration. Constraints were established to ensure that user demands are met and equipment is utilized efficiently. An enhanced genetic algorithm was employed to minimize daily operational costs， with optimized scheduling performed for typical days across different seasons. The results demonstrate that the multi?energy supply system is economically viable， with operational costs recorded at 3 480.17， 4 752.25， 3 168.23， and 4 464.02 yuan for typical days in spring， summer， autumn， and winter， respectively. The system’s energy comprehensive utilization efficiencies reached 65.10%， 57.82%， 67.25%， and 70.12% for the same periods. Although the energy utilization efficiency is slightly lower compared to traditional natural gas?fueled CCHP system， the proposed system offers cost advantages and aligns with China’s “more coal and less gas” energy supply situation.

The effect of different machining parameters on the milling force of high?entropy alloy（HEA） was studied through the simulation analysis of 3D slot milling and side milling. The orthogonal and single?factor milling experiments of FeCoNiCrAl_0.1， FeCoNiCrAl_0.5， and FeCoNiCrMo_0.1 as?cast HEAs were designed. By measuring the milling force in the experiments， the effects of different machining methods， milling parameters， tools， and elements of HEAs on the milling force were explored. The results showed that with the increase of milling speed and the decrease of depth and feed speed， the milling force of slot milling and side milling decreases， and the tangential force and normal force of slot milling are 404.30% and 761.06% higher than that of side milling under the same machining parameters， respectively. High?entropy alloys with high Al content have higher milling forces. The milling force of HMX tool is smaller than that of SGS tool， with an average reduction of 10.6% in the milling force. The experimental results provide a theoretical reference for the efficient machining of HEAs.

Given that it is difficult to accurately diagnose rolling bearing faults in a strong noise environment， a deep residual shrinkage network is proposed with pooling fusion for rolling bearing fault diagnosis. Firstly， the proposed method employs residual connections to avoid the risk of gradient vanishing or explosion due to excessive network depth. Then， the approach uses multi?scale pooling feature fusion to extract more comprehensive local features from vibration signals， and an attention mechanism to automatically derive the optimal threshold of a soft threshold function. Finally， the network is trained through labeled data to achieve the accurate fault diagnosis of rolling bearings in a strong noise environment. Experiment results show that the deep residual shrinkage network based on pooling fusion can outperform the traditional models in diagnosing rolling bearing faults under noise conditions with different SNRs.

The surface quality of riveting holes on aircraft skin， tail and other components is crucial to the overall assembly performance of an aircraft. Currently， most riveting hole defect detection relies on the traditional manual methods， which are prone to missing detection. An improved detection method based on YOLOv8 for surface defect detection of riveting holes was proposed. The conventional convolution was replaced by deformable convolution to solve the problem of the fixed receptive field shape in feature extraction. The SimAM attention mechanism was embedded in order to enhance the recognition ability of the network under low contrast between the background and targets. The CIoU loss function was replaced by the WIoU bounding box regression loss function to reduce the impact of low?quality images during model training and improve the robustness and generalization ability of the model. To verify the performance of the improved model， 6061 aluminum alloy plates with riveting holes were used as a substitute for aircraft skin in the detection process. Experimental results demonstrated that the improved model achieved mAP_0.5 and accuracy of 0.918 and 0.920 on the riveting hole test set， which represents an improvement of 24.1% and 25.3% compared to the original model.

With the increasing span of steel-UHPC lightweight composite girders and the horizontal shear force in girders， the design of shear performance of headed studs in the lightweight composite girders becomes increasingly critical. The shear performance of headed studs between steel and UHPC with small coarse aggregates is studied via standard push?out specimens. The load?slippage behavior and failure mode of standard specimens are revealed， and the calculation methods for shear stiffness and capacity of single headed stud are proposed. Test results show that the failure mode of specimens is that the roots of headed studs are cut off. The failure slippage range of the push?out specimen is between 0.5~1.2 mm， and the ultimate shear strength of the single headed stud is 400.7 MPa. Theoretical calculation results show that according to existing design specifications， the shear stiffness is designed safely at the half failure load or the slippage of 0.2 mm. The design shear capacity of the single headed stud is calculated in the range of 95.51~129.20 kN， which was only 62.7%~84.8% of the test results. Therefore， when designing headed studs for UHPC deck panels with small coarse aggregates， the shear stiffness should consider a reduction coefficient of 0.57~0.73， and the shear capacity should not be reduced.

To response to high difficulty and cost in the blasting operation for cutting vertical slots in sublevel stope， a mine in Shandong as the engineering background is used to conduct research on the blasting mechanism of the “pin”?shaped slotting without raise in sublevel stope. The LS-DYNA numerical simulation method is employed to analyze the differences between the “pin”?shaped without raise cutting blasting and the VCR method cutting blasting. The study reveals the multi?free?surface effect of the “pin”?shaped blasting for slot cutting. The numerical results show that under the combined action of lateral and bottom free surfaces， the blasting stress distribution of the cylindrical blastholes in the “pin”?shaped without raise cutting blasting method is more uniform along their length， enhancing the rock?breaking capability at the free surfaces， reducing damage to the unexploded area， and improving the utilization of blasting energy. Through comprehensive comparison， the overbreak rate of the “pin”?shaped without raise cutting blasting method is reduced by 23.2%， the effective rock?breaking stress is increased by 12 MPa， and the vibration values of the surrounding rock in the unexploded area are reduced by 16.7%， significantly improving the blasting effect of slot cutting in sublevel stope.

A binocular stereo matching algorithm that integrates large?scale window information and Manhattan distance is proposed to address the low matching accuracy of traditional methods， such as AD-Census， in areas with low or repeated textures. The algorithm first uses an improved SAD cost and multi?gray threshold Census cost to calculate the fusion cost， and assigns weights based on the Manhattan distance between neighboring pixels and the center point to reduce the influence of edge pixels on the cost. The algorithm also screens and filters the difference information extracted from large scale windows to improve the accuracy in areas with repeated textures and low gray similarities. Compared to traditional AD-Census algorithm， the proposed algorithm reduces the false matching rate by approximately 18%. Furthermore， the algorithm has been transplanted to the GPU， allowing it to run 1~2 orders of magnitude faster on images with different scale resolutions， thus meeting the demands of quick and accurate binocular stereo matching.

To study the effect of pore characteristics on filtration performance of filter bag， the maximum pore size and mean pore size of thirteen kinds of samples are determined by bubble?point method， mainly including needled filter bag， spunlace filter bag and membrane covered filter bag. In addition， the pore distribution characteristics of needled filter bag and spunlace filter bag are also analyzed by a combination of scanning electron microscope and Image-pro plus 6.0. The results show that the maximum pore size and mean pore size of needled filter bag are larger than those of the spunlace filter bag， and the pore size of the membrane covered filter bag is the lowest one. The maximum pore size and mean pore size of the internal pore throat have an effect on the air permeability， filtration efficiency and pressure drop of filter bags. Air permeability increases with the increase of pore size. The filtration efficiency increases with the decrease of pore size. Additionally， the pore distribution of spunlace filter bag is more concentrated and uniform than other kinds of filter bags， which has a certain degree of influence on the improvement of filtration efficiency of filter materials.

The lower flammability limits of H₂/CH₄ mixtures over a wide range of H₂ fractions are systematic experimentally studied at various initial temperatures. The effect of initial temperatures on the near?limit flames of H₂/CH₄ mixtures is insightfully discussed. In addition， the estimations of the lower flammability limits of H₂/CH₄ mixtures is also calculated using Le Chatelier rule and Kondo’s correlation. The results show that the lower flammability limits of H₂/CH₄ mixtures decrease with the growth of H₂ fractions under all measured initial temperatures. The greater declines of lower flammability limits are presented at higher initial temperatures. Moreover， the lower flammability limits of H₂/CH₄ mixtures almost linearly decrease with the increase of initial temperatures. Under the higher initial temperatures， the dominant promotion of H+O₂=O+OH on the near-limit flames of H₂/CH₄ mixtures is enhanced by the significant increase of H radicals. The comparison between the experimental data and calculation results shows that Le Chatelier rule presents satisfactory applications on estimating lower flammability limits of H₂/CH₄ mixtures based on the accurate measurements. Kondo’s correlation provides the greater accurate predictions on the linear temperature dependence of the lower flammability limits of mixtures with lower H₂ fractions. However， the considerable distinctions between experimental data and calculation results using Kondo’s correlation at higher H₂ fractions under elevated initial temperatures are attributed to the significant variation of the sensitivity of dominant reactions and the mole fractions of vital radicals.

In response to the challenges of low bearing capacity and complex construction procedures in traditional pipe roofing method， a novel pipe roofing method， namely small?diameter pipe roofing method is proposed， which is applied in the construction of the Heping South Street Station of Shenyang Metro. A three?dimensional finite difference model for this project is established， to investigate the ground deformation characteristics during the construction. The influence of pipe?roof parameters on four indicators： maximum surface settlement at the mid?span after the station construction， ground disturbance caused by pipe roofing construction， pipe roofing stiffness， and cost is clarified. Finally， the pipe?roof parameters are optimized using fuzzy mathematics theory. The research results show that when constructing a subway station using the small?diameter pipe roofing method， the stage of removing the pilot tunnel wall and constructing the roof is the critical for controlling ground deformation. The most critical parameter affecting the above four indicators is the steel tube diameter. The optimized combination of parameters is steel tube diameter of 400 mm， steel tube center distance of 800 mm， steel tube wall thickness of 10 mm， concrete strength grade inside the tube is C40， and the steel tube strength grade is Q390.

Combining the electroencephalogram， eye tracking， and subjective questionnaire survey methods， how social embodied agent type and manifestation could affect the neurophysiological and psychological responses of users， and then affect agent acceptance was explored. The results indicated that users feel more unfamiliar with agents with animal appearance， which drive unconscious perceptual processing and encoding， and take up more attention resources. Users perceive agents with human appearance to be more independently conscious than agents with animal appearance. Agents with animal appearance and agents manifested with real human or animal appearances are more attractive or complicated to users， which require users to invest more attention resources. Agents with real female appearance arouse the highest agent’s likeability due to their resemblance to users’ own images， while agents with real pet dog appearance evoke the lowest agent’s likeability. User acceptance of agents with real female appearance is the highest， followed by agents with cartoon female appearance and cartoon pet dog appearance， while agents with real pet dog appearance is the lowest. The research results provide guidance for the design of social embodied agents as virtual social companions.

The pricing and refunding decision models for a pricing and refunding problem of an omnichannel retailer with “buy?online?and?deliver?from?store” （BODS）， “buy?online?and?return?

online” （BORO） and “buy in?store” （BS） channels are established under the full and partial refunding strategies. The retailer’ s optimal pricing and refunding strategies are determined. Furthermore， the effects of factors such as the distribution range， the proportion of consumers across channels， and the salvage value of products returned online or in?store on optimal pricing and refunding decisions as well as on the optimal profits are analyzed， and the conditions under which the retailer should implement a full or partial refunding strategy are explored. Finally， numerical studies are verified. The results show that when the salvage value of products returned online is less than or equal to that of products returned to the store， the partial refunding strategy is better. Conversely， the proportion of online consumers who choose BORO within the store distribution range will affect the retailer’s refunding strategies.