查看更多>>摘要：Electric vehicles have been rapidly developing worldwide due to the use of new energy.Howev-er,at the same time,serious traffic accidents caused by driver fatigue in emergency situations have also drawn widespread attention.The lack of datasets in real vehicle test environments has always been a bottleneck in the research of driver fatigue in electric vehicles.Therefore,this study establi-shes a dataset from real vehicle test,applies the Bayesian optimization support vector machine(BOA-SVM)algorithm to take features of electromyography(EMG)and electrocardiography(ECG)signals as input and develop an early warning model for driving fatigue detection.Firstly,the driver's EMG and ECG signals are collected through real vehicle testing experiments and then combined with the driver's subjective fatigue evaluation scores to establish the dataset.Secondly,the study establishes a driver fatigue early warning model for emergency situations.Time-domain and frequency-domain features are extracted from the EMG signals.Principal component analysis(PCA)is applied for dimensionality reduction of these features.The experimental results show that based on the input of dimensionality reduced EMG features and ECG features,the BOA-SVM algorithm achieved an accuracy of 94.4%in classification.
查看更多>>摘要：Due to the issue of long-horizon,a substantial number of visits to the state space is required during the exploration phase of reinforcement learning(RL)to gather valuable information.Addi-tionally,due to the challenge posed by sparse rewards,the planning phase of reinforcement learning consumes a considerable amount of time on repetitive and unproductive tasks before adequately ac-cessing sparse reward signals.To address these challenges,this work proposes a space partitioning and reverse merging(SPaRM)framework based on reward-free exploration(RFE).The framework consists of two parts:the space partitioning module and the reverse merging module.The former module partitions the entire state space into a specific number of subspaces to expedite the explora-tion phase.This work establishes its theoretical sample complexity lower bound.The latter module starts planning in reverse from near the target and gradually extends to the starting state,as opposed to the conventional practice of starting at the beginning.This facilitates the early involvement of sparse rewards at the target in the policy update process.This work designs two experimental envi-ronments:a complex maze and a set of randomly generated maps.Compared with two state-of-the-art(SOTA)algorithms,experimental results validate the effectiveness and superior performance of the proposed algorithm.
查看更多>>摘要：In the structure of double-nut ball-screw mechanism(BSM),the contact angle of the ball-screw determines the relative positional relationship between the balls and the screw as well as the nut.The contact angle is related to geometrical parameters of the ball,the screw and the nut,which are also affected by the running status and the preload of the BSM.Considering the effect of the gy-roscopic moment on the ball in the raceway,the dynamic model of the ball in space is established under different speeds and different preloads of the BSM.By simulation of the dynamic model of the ball in space,the changing regularity of the contact angle,the helix angle,the drag torque and the mechanical efficiency of the BSM can be obtained under different speeds and different preloads.The results show that there is a nonlinear gradient relationship between contact angle,helix angle,the drag torque,the mechanical efficiency and the speeds of the ball-screw under different preloads.The contact angle is the key factor to affect the drag torque of the BSM.Through the analysis,it is found that establishing the ball dynamic model in space can better study the precision degradation law of the ball screw.
查看更多>>摘要：Traffic flow forecasting plays a crucial role and is the key technology to realize dynamic traffic guidance and active traffic control in intelligent traffic systems(ITS).Aiming at the complex local and global spatial-temporal dynamic characteristics of traffic flow,this paper proposes a new traffic flow forecasting model spatial-temporal attention graph neural network(STA-GNN)by combining at-tention mechanism(AM)and spatial-temporal convolutional network.The model learns the hidden dynamic local spatial correlations of the traffic network by combining the dynamic adjacency matrix constructed by the graph learning layer with the graph convolutional network(GCN).The local tem-poral correlations of traffic flow at different scales are extracted by stacking multiple convolutional kernels in temporal convolutional network(TCN).And the global spatial-temporal dependencies of long-time sequences of traffic flow are captured by the spatial-temporal attention mechanism(STAtt),which enhances the global spatial-temporal modeling and the representational ability of model.The experimental results on two datasets,METR-LA and PEMS-BAY,show the proposed STA-GNN model outperforms the common baseline models in forecasting accuracy.
查看更多>>摘要：In response to the problem of inter-carrier interference(ICI)and inter-subband interference(IBI)in the received signals of universal filtered multi-carrier(UFMC)systems,a novel interfer-ence suppression design scheme applying the method of complex weighted matrix inter-leaving map-ping(CWMIM)is proposed on the basis of the existing suppression scheme of conjugate weighted butterfly interleaving mapping(CWBIM).The proposed scheme performs matrix interleaving map-ping on the transmitted signal,which not only improves the carrier interference ratio(CIR)of the received signal by combining the original IBI and ICI terms,but also further inhibits the probability of burst error in the received signal.Meanwhile,the scheme can further decrease the impact of phase rotation errors in the received signal by increasing the number of rotation factors.Theoretical analysis and simulation results demonstrate that compared with CWBIM-UFMC,the proposed CWMIM-UFMC can obtain more effective ICI and IBI suppression and better system bit error rate(BER)performance with only a little bit increase in computational complexity.
查看更多>>摘要：The performance of deep learning models is heavily reliant on the quality and quantity of train-ing data.Insufficient training data will lead to overfitting.However,in the task of alert-situation text classification,it is usually difficult to obtain a large amount of training data.This paper proposes a text data augmentation method based on masked language model(MLM),aiming to enhance the generalization capability of deep learning models by expanding the training data.The method em-ploys a Mask strategy to randomly conceal words in the text,effectively leveraging contextual infor-mation to predict and replace masked words based on MLM,thereby generating new training data.Three Mask strategies of character level,word level and N-gram are designed,and the performance of each Mask strategy under different Mask ratios is analyzed and studied.The experimental results show that the performance of the word-level Mask strategy is better than the traditional data augmen-tation method.
查看更多>>摘要：During electric vehicle(EV)-assisted grid frequency modulation,inconsistent state of charge(SOC)among EVs can result in overcharging and discharging of the batteries,affecting the stability of the electrical system.As a solution,this paper proposes a priority-based frequency regulation strategy for EVs.Firstly,models for the primary and secondary frequency regulation of EV-assisted power grids are established.Secondly,a consensus algorithm is used to construct a distributed com-munication system for EVs.Target SOC values are used to obtain a local frequency regulation priori-ty list.The list is used in an optimal control plan allowing individual EVs to participate in frequency regulation.Finally,a simulation of this strategy under several scenarios is conducted.The results indicate that the strategy ensures uniform SOC among the participating group of EVs,thereby avoi-ding overcharging and discharging of their batteries.It also reduces frequency fluctuations in the electrical system,making the system more robust compared with the frequency regulation strategy that is not priority-based.
查看更多>>摘要：In this paper,a multi-strategy improved coati optimization algorithm(MICOA)for engineering applications is proposed to improve the performance of the coati optimization algorithm(COA)in terms of convergence speed and convergence accuracy.First,a chaotic mapping is applied to initial-ize the population in order to improve the quality of the population and thus the convergence speed of the algorithm.Second,the prey's position is improved during the prey-hunting phase.Then,the COA is combined with the particle swarm optimization(PSO)and the golden sine algorithm(Gold-SA),and the position is updated with probabilities to avoid local extremes.Finally,a population decreasing strategy is applied as a way to improve the performance of the algorithm in a comprehen-sive approach.The paper compares the proposed algorithm MICOA with 7 well-known meta-heuristic optimization algorithms and evaluates the algorithm in 23 test functions as well as engineering appli-cation.Experimental results show that the MICOA proposed in this paper has good effectiveness and superiority,and has a strong competitiveness compared with the comparison algorithms.
查看更多>>摘要：Intelligent reflecting surface(IRS)can efficiently improve the performance of wireless commu-nication networks by intelligently reconfiguring the wireless propagation environment.Recently,IRS has been integrated with cognitive radio(CR)network in order to improve the resource utilization of communication systems.It is a challenging issue for IRS-assisted CR networks to improve the rate performance of the secondary user(SU)through the rational design of IRS passive beamforming while limiting the interference to the primary network.This paper investigates the optimization of downlink rate of SU in a double-IRS-assisted CR network.The achievable rate is maximized by jointly optimizing the active beamforming vector at the secondary transmitter(SU-TX)and the coop-eratively passive reflective beamforming at the two distributed IRSs.To solve the proposed non-con-vex joint optimization problem,the alternating optimization(AO)and semidefinite relaxation(SDR)techniques are then adopted to iteratively optimize the two variables.Numerical results vali-date that the proposed double-IRS assisted system can significantly improve the performance of the CR network compared with the existing single-IRS assisted CR system.
查看更多>>摘要：In practical applications,different power companies are unwilling to share personal transformer data with each other due to data privacy.Faced with such a data isolation scenario,the centralized learning method is difficult to be used to solve the problem of transformer fault diagnosis.In recent years,the emergence of federated learning(FL)has provided a secure and distributed learning framework.However,the unbalanced data from multiple participants may reduce the overall per-formance of FL,while an untrusted central server will threaten the data privacy and security of cli-ents.Thus,a fault diagnosis of intelligent distribution system method based on privacy-enhanced FL is proposed.Firstly,a globally shared dataset is established to effectively alleviate the impact of un-balanced data on the performance of the FedAvg algorithm.Then,Gaussian random noise is intro-duced during the parameter uploading process to further reduce the risk of data privacy leakage.Fi-nally,the effectiveness and superiority of the proposed method are verified through extensive experi-ments.
万方数据
维普