Addressing the challenges inherent in passive unmanned search and rescue missions at sea,including difficulties in target identification,broad search areas,and slow route planning,a strategic process was introduced for maritime search and rescue area planning and a route planning model specifically designed for passive unmanned missions.By thoroughly understanding the emergency response operations at sea and the specific needs for route planning,an optimal routing model have been developed considering factors such as the efficiency of search and rescue area coverage and the cost of rescue routes.The objective function is constructed within these constraints and solved using the whale optimization algorithm.The validity of the model is confirmed through designated scenario experiments,indicating that our proposed model for maritime passive unmanned search and rescue route planning is capable of swiftly identifying the search and rescue area and efficiently discovering a route with reduced costs.
To address the problem of low cache hit ratio in edge nodes for privacy-preserving in the Internet of Vehicles(IoV),a deep deterministic policy gradient caching(DDPGC)method was proposed.Firstly,a taxi certified by a trusted authority acted as a second-level caching edge node to acquire hotspot data and store it in the local cache.It then broadcasted this information to the neighboring service requesting vehicles(SRV).SRVs cached the broadcasted data locally and search for service requests in the order of priority of local cache,taxi,and cloud server when such requests arise.Secondly,a neural network was deployed in taxis and SRV to maximize the caching benefit through deep reinforcement learning for decision replacement of their cached data.Finally,when SRV were located in vehicle sparsity and could not obtain request data from neighboring vehicles,a combination of k-anonymity and random response perturbation mechanisms generated anonymity sets to send requests to cloud servers in an anonymous manner to obtain services while protecting user location privacy.Simulation experimental results show that DDPGC can effectively improve the vehicle cache hit ratio,reduce the frequency of SRV interaction with the cloud server,and effectively protect user privacy security.
Aiming at the problems of traditional image manipulation detection methods,such as fuzzy boundaries,single scale of extracted features,and ignoring background information,this paper proposes an image manipulation detection method based on multi-scale context-aware and boundary-guided.First,spatial details and base features of manipulated images are extracted using an improved pyramid vision transformer.Second,information related to the edge of the falsified region is explored by an edge context-aware module to generate an edge prediction map.Again,the edge guidance module is utilized to highlight the key channels in the extracted features and reduce the interference of redundant channels.Then,the rich contextual information of the manipulated region is learned from multiple sensory fields through the multi-scale context-aware module.Finally,the feature fusion module is utilized to accurately segment the manipulated region by focusing alternately on the foreground and background of the manipulated images.Comparing this paper's method quantitatively and qualitatively on five commonly used public image manipulation detection datasets,the experimental results show that this paper's method can effectively detect manipulated regions and outperforms other methods.
Aiming at the limitations of single-dimensional car-following model to describe vehicle car-following behavior in local multi-vehicle environment,the mechanism of vehicles in adjacent lane influencing the subject vehicle car-following behavior is explored,and a vehicle car-following model more suitable for local multi-vehicle environment is attempted to be established.The driving behavior variable reflecting the influence of vehicles in adjacent lane was determined by correlation analysis.Vortisch indicator of similarity(VIS)was used to characterize the influence of vehicles in adjacent lane on car-following behavior.Chi-square independent test and kernel density curve were used to determine the VIS demarcation threshold reflecting whether the influence is significant.Recursive feature elimination was used to screen the variables related to car-following samples significantly affected by vehicles in adjacent lane.The influence mechanism of variables was determined according to the statistical analysis results.Based on the mechanism proposed,the car-following model suitable for local multi-vehicle environment was constructed and its prediction effect was evaluated.Results show that VIS between the speed of subject vehicle and following vehicle in adjacent lane can characterize the influence and the VIS threshold is 0.668.It is concluded that the attention mechanism and memory effect can explainthe influence mechanism of car-following behavior in multi-vehicle environment.Considering the attention mechanism and memoryeffect,RMSE decrease by 65%in full velocity difference model(FVD)model and 62%in intelligent driver model(IDM)model,MAE decrease by 65%and 59%and R2 increased by 180%and 288%respectively,which proved the rationality of the attention mechanism and memory effect in explaining the car following behavior of subject vehicle in local multi-vehicle environment.
Traffic risk is an important source of road traffic accidents,and effective prediction of road accident risk propagation after an accident is of great significance to prevent the occurrence of secondary accidents.Therefore,this paper establishes a road accident risk propagation model based on macroscopic traffic flow energy consumption.Firstly,the influence of lane occupancy and lane change on traffic flow after accident is analyzed,and then a macro-traffic flow model is constructed under the effect of post-accident risk.Secondly,combined with the definition of traffic flow energy dissipation,the whole vehicle movement is abstracted as fluid movement process,and the traffic flow energy consumption model under the action of post-accident risk is proposed to quantify the accident risk.Finally,according to the temporal and spatial correlation of risk transmission,a calculation model of risk influence range and duration is proposed to describe the process of risk transmission.In order to verify the effectiveness of the risk propagation model proposed in this paper,numerical simulation and simulation verification of traffic flow operation under different initial traffic flow densities were carried out using MATLAB and VISSIM respectively.The experimental results show that:the greater the initial density of the road section when the accident occurs,the wider the impact range of the risk and the longer the duration.The absolute value of the prediction deviation rate of this model is less than 8%,which has good prediction effect.
In addressing the issue of relatively low accuracy in node classification tasks on heterophily graphs using methods such as MLP and GCN,a Graph Neural Network based on Similarity Random Walk Aggregation(SRW-GNN)was proposed.To address the impact of heterophily on node embeddings,SRW-GNN employs the similarity between nodes as probabilities for conducting random walks.The sampled paths serve as the neighborhood,enabling the model to gather more homophily-based information.To address the issue of insensitivity to node order in most existing graph neural network(GNN)aggregators,a path aggregator based on recurrent neural networks(RNNs)was introduced to simultaneously extract features and order information of each node in the path.Furthermore,nodes exhibit varying preferences for different paths.To adaptively learn the importance of different paths in node encoding,an attention mechanism was employed to dynamically adjust the contributions of each path to the final embedding.Experimental results on several commonly used heterophily graph datasets demonstrate that the proposed SRW-GNN method achieves significantly higher accuracy compared to the methods such as MLP,GCN,H2GCN,HOG-GCN,validating its effectiveness in heterophily graph node classification tasks.
To satisfy the testing requirements for intelligent vehicles and pedestrians interaction under urban conditions， a scenario generation method that comprehensively considers the appearance frequency of scenarios in the real world and their challenges to intelligent vehicles performance is proposed. First， the original scenarios are extracted from the natural driving dataset. Then a critical scenario extraction method based on importance sampling theory is designed to extract essential scenarios from the original scenarios according to the accelerated test requirement， and pedestrian crossing road scenarios based on natural driving data are constructed. Finally， comparing the distribution of essential scenarios and original scenarios，the results show that this method can effectively screen out scenarios that may pose challenges to intelligent vehicles safety and it also realizes accelerated testing while retaining the statistical characteristics of test scenarios.
In order to meet the energy absorption requirement of the landing buffer,this paper starts with the buffer filling material of the buffer,draws lessons from the Kelvin structure and spiral structure,and uses the engineering bionic principle to design and establish two kinds of structure models:footballene bionic structure and multi-helix bionic structure.Considering the requirement of energy absorption and reuse of the buffer structure of the lander,the biomimetic structure samples were prepared by NiTi alloy with shape memory effect and were made by additive manufacturing technology.The mechanical properties,energy absorption ability and recoverability of the samples were analyzed and verified by simulation and experiments.The results show that the accuracy of the numerical simulation is verified by comparing the force-displacement curves of the simulation test and the isometric static pressure test,in which the multi-helix bionic structure has better mechanical properties,and its maximum energy absorption is 3 096.23 J;the recovery rates of the two structures are as high as 98.02%and 97.12%,respectively,and the recovery rate of footballene bionic structure is slightly higher.In this study,the bionic buffer structure of the leg lander is prepared by the method of adding materials,which provides a reference for the bionic design of the buffer structure of the lander.
To satisfy the massive computational requirement of Convolutional Neural Networks,various Domain-Specific Architecture based accelerators have been deployed in large-scale systems.While improving the performance significantly,the high integration of the accelerator makes it much more susceptible to soft-error,which will be propagated and amplified layer by layer during the execution of CNN,finally disturbing the decision of CNN and leading to catastrophic consequences.CNNs have been increasingly deployed in security-critical areas,requiring more attention to reliable execution.Although the classical fault-tolerant approaches are error-effective,the performance/energy overheads introduced are non-negligible,which is the opposite of CNN accelerator design philosophy.In this article,we leverage CNN's intrinsic tolerance for minor errors and the similarity of filters within a layer to explore the Approximate Fault Tolerance opportunities for CNN accelerator fault tolerance overhead reduction.By gathering the filters into several check groups by clustering to perform an inexact check while ensuring that serious errors are mitigated,our approximate fault tolerance design can reduce fault tolerance overhead significantly.Furthermore,we remap the filters to match the checking process and the dataflow of systolic array,which can satisfy the real-time checking demands of CNN.Experimental results exhibit that our approach can reduce 73.39%performance degradation of baseline DMR.
Aiming at the defects of traditional density-based uncertain clustering algorithm， such as parameter sensitivity and poor clustering results for complex manifold uncertain data sets， this paper proposes a new uncertain data density peak clustering algorithm based on JS divergence （UDPC-JS）. The algorithm first uses the uncertain natural neighborhood density factor defined by uncertain natural neighbors to remove noise points; secondly， the local density of uncertain data objects is calculated by combining uncertain natural neighbors and JS divergence. Then， the initial clustering center of uncertain data sets is found by combining the idea of representative points， and the distance based on JS divergence and graph is defined between the initial clustering centers. Then， the local density calculated based on uncertain natural neighbors and JS divergence and the newly defined distance based on JS divergence and graph between the initial clustering centers are used to construct the decision graph on the initial clustering center， and the final clustering center is selected according to the decision graph. Finally， the unassigned uncertain data objects are assigned to the cluster where their initial clustering centers are located. The experimental results show that the algorithm has better clustering effect and accuracy than the comparison algorithm and has greater advantages in dealing with uncertain data sets of complex manifolds.
Understanding passenger travel patterns is helpful to the allocation of passenger resources in urban rail transit. Based on the smart card data of urban rail transit， this paper proposes a method to identify travel patterns by modeling the spatio-temporal sequences of individuals. First， all stations visited by a passenger individual were extracted， and the similarity of stations was calculated in terms of inter-station travel frequency， inter-station distance， and the station activity duration， thus the main spatial activity areas of this individual were classified using a hierarchical clustering algorithm. Then， the spatio-temporal sequence was inferred based on the travel order of the individual， which is a set of discrete values characterizing the spatio-temporal state. PCA-KL and K-Means++ were used to extract the similarity sequence structure to identify passenger travel patterns. Finally， using one-month smartcard data as an example to identify passenger travel patterns for Xi’an rail transit. The results show that the complex passenger flow has five travel patterns， among which three typical travel patterns are macroscopic commuter travel behavior， accounting for 79% of the total passenger flow. Thus， the pattern identified based on the similarity of individuals’ travel spatiotemporal sequences fully reflects the particularity and universality of the research method and it is highly operable for different cities.
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.
Most of the existing sentiment analysis methods based on commodity comments texts seldom consider the aspect features,and the related analysis models can't consider both the long-term context dependence and the local text features,thus affecting the accuracy of sentiment analysis.To solve the problem,a sentiment analysis method based on bidirectional gated recurrent network(BiGRU)and capsule network is proposed.Firstly,the method uses the approach based on the word frequency statistics to extract the aspect features from the comment texts,and integrate them into the word vector representation to effectively improve the expression ability of the word vector.Then,the BiGRU network is used to extract the long-term context features of the texts,and Capsule Network is used to extract local features of the comment texts,thus realizing high-precision text sentiment analysis based on aspect.The experimental results on real datasets show that the proposed method is superior to bidirectional long-short term memory(BiLSTM),CNN-LSTM,TextCNN and other sentiment classification models in Accuracy,Precision,Recall and F1 score.
The distributed trajectory tracking problem for multiple UAVs under unknown external disturbance is investigated.A new nonlinear robust trajectory tracking strategy for multiple UAVs is proposed.The dynamic model for the UAVs' formation is illustrated in the Tangent frame.For the trajectory tracking problem,a robust formation tracking control strategy based on robust integral of the signum of error(RISE)is designed to compensate for the effects of unknown external disturbances,which improves the robustness of the UAVs' formation system.Based on the Lyapunov stability analysis,it is proved that the global asymptotic convergence of the coordination errors and the semi-global asymptotic convergence of the UAVs' position tracking errors are achieved.The proposed formation control strategy is validated via real-time experiments that are performed on the self-build UAVs' formation flight control testbed.Flights without wind disturbance and flights with disturbance are all performed.The performance comparison experiment with the conventional sliding mode control algorithm is performed.The experimental results show that the designed control strategy can achieve good coordinated trajectory tracking of multiple UAVs,and has better control performance compared with normal sliding mode control law.
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.
In order to investigate the effects of loading and high temperature on the uniaxial compressive properties of desert sand concrete （DSC）， the uniaxial compression experiment of DSC subjected to different loading and temperatures were carried out to obtain the stress-strain curves. The influences of loading level， temperature and cooling methods on the mass loss rate， ultrasonic velocity and axial compressive properties of DSC were analyzed. Experimental results showed that the mass loss rate of DSC gradually increased with the temperature. The uniaxial compressive strength and modulus of elasticity of DSC declined， peak strain increased greatly， and the stress-strain curve gradually became flat. The "pseudoplastic plateau" near uniaxial compressive strength of DSC stress-strain curve became more obvious. Taking into account of temperature and loading level， the stress-strain model was established to simulate the mechanical properties of DSC on the basis of two-stage constitutive model， which provided technical support for performance evaluation of DSC after high temperature.
The main structure and characteristic parameters affecting the working performance of vertical screw stirring mill were analyzed and studied， and the selection method of key working parameters was proposed. Based on the discrete element method， the simulation model of vertical screw stirring mill was established， and the effects of spindle speed， agitator lead and grinding medium size distribution on grinding performance were analyzed. The comprehensive grinding performance index was put forward， orthogonal tests were carried out for each parameter， and the quantitative value of grinding effect was obtained， and then the optimal working parameter combination under specific weight coefficient was obtained， which provided a reference method for the optimal design of the mill.
Mass spectrometry is commonly used for disease prevention and diagnosis， but the large number of mass spectrometry data features and the wide variation of features among different diseases make the task of multi-disease diagnosis complex and difficult. To solve the above problems， this paper proposes the generative adversarial autoencoder integrated voting algorithm msDAGVote based on mass spectrometry data. The msDAGVote feature extraction framework uses a dual autoencoder-based generative adversarial network， and after the network has been trained by mass spectrometry data， the generator sub-network is used for feature construction. Evaluated using mass spectrometry datasets of 10 different disease types， the experimental data show that msDAGVote extracts better features than comparative method， significantly reduces the number of features required for classification while providing excellent diagnostic power for disease classification， with classification AUC over 0.98 on six datasets and 0.87 on the remaining challenging datasets.
Based on the seismic design concept of recoverable function,a replaceable splicing joint with friction energy dissipation components was proposed. The joint was composed of H-steel column,cantilever section,H-steel beam,replaceable flange connection plate and friction hinge. The finite element model of the joint was established by finite element software ABAQUS,and the pseudo-static analysis was carried out. By changing the strength and thickness of the flange connection plate,the depth and length of the weakening zone,and the diameter and number of the rotating bolts,the effects of different design parameters on the hysteretic performance,bearing capacity,stiffness and displacement ductility of the joint were investigated. The results showed that the bearing capacity,stiffness and total energy consumption of the replaceable splicing joint increased with the increase of the strength of the flange connection plate and the friction force of the friction hinge,and the ductility of the joints with LYP160 steel in flange connection plate was significantly higher than that of the joints with Q235 steel. The friction energy dissipation of friction hinge increased with the increase of friction force,and decreased with the increase of the strength of flange connection plate.
In response to the problem of heavy adhesion and high resistance of the rice direct seeder skateboard in the disturbed and saturated paddy field,the non-smooth surface of the loach body was observed microscopically;Using Fluent software for simulation and analysis,the principle of reducing viscosity and resistance on the non-smooth surface has been revealed;The Box Behnken response surface method was used to design the experiment and obtain the regression equations of the total resistance FB of the non-smooth surface of the groove with the groove width w,groove depth h,and inflow velocity v.The optimization analysis results showed that the maximum drag reduction rate reached 10.052%when w was 4.5 mm,h was 4 mm,and v was 1.25 m/s;Based on the optimal parameters,a biomimetic sliding plate for rice direct seeding machine was designed,and indoor paddy soil tank experiments were conducted,with a drag reduction rate of 10.23%ultimately.This study can provide new ideas for the development of viscosity reduction and resistance reduction technology for paddy field soil contacting components.
