Auditory distance perception is pivotal in recognizing potential dangers，determining sound source locations，daily activities，interpersonal communication，music and art performances，as well as the processes of learning and memorizing. This study integrated subjective evaluations with event⁃related potential (ERP) analysis，employing behavioral experiments and electroencephalography (EEG) physiological measurements to explore the phenomena and the underlying mechanisms of auditory distance perception. Subjective evaluation results indicated that participants' distance perception was more accurate at closer ranges，whereas accuracy decreased at greater distances，displaying an underestimation consistent with a power function fit. Behavioral results demonstrated that compared to the near deviant stimuli，participants exhibited significantly higher error rates and more reaction time in response to the far deviant stimuli. In addition，ERP results showed the presence of the P1⁃N1⁃P2 complex. In the no⁃attention condition，the P3a component appeared only for the more pronounced near deviant stimuli; however，under the attention condition，it was present for both near and far deviant stimuli. The P3b component was significant only under the attention condition，and as the stimuli intensity decreased and task difficulty increased，its latency increased and amplitude decreased. ERP imaging also revealed hemisphere⁃specific responses in the brain，with the right hemisphere being more sensitive to distance⁃related auditory stimuli.
The sound zone control system propagates different audio information to various regions in a sound field. Due to the directionality of sound sources,the sound power radiated in different directions varies. In this paper,the influence of the sound source directivity on the performance of sound zone control is investigated for the application in automotive cabins. Firstly,simulations of the sound zone control are conducted based on the free⁃field model of the sound source. The contributions of sound sources located at different positions to the sound energy in the bright and dark zones are analyzed for cases where the front and rear zones of the cabin are bright zones,respectively. The impact of the sound source directivity on the sound zone control performance is also investigated and a qualitative explanation is given. Secondly,the compound source with different directivity is designed by optimizing the amplitude ratio and phase difference between two enclosed speakers. Finally,experiments are conducted in an anechoic chamber to compare the effects of the sound zone control obtained by using compound sources with different directivities,and the experimental results are compared with those obtained from simulations. It is shown that the directivity of the sound source significantly influences the acoustic contrast when the front zone of the cabin is the bright zone. By adjusting the directivity of the compound source,the sound zone control performance can exceed that achieved with monopole source. However,when the rear zone of the cabin is the bright zone,no significant differences in the control performance are observed across sound sources with different directivities.
This paper presents an innovative approach for accurately locating buried micro⁃cracks in petrochemical pipelines by integrating a fusion feature matrix with a convolutional neural network (CNN) model. The study addresses the challenge of detecting and localizing micro⁃cracks within pipeline structures,which are more difficult to identify and pose greater risks than surface⁃breaking cracks do.The methodology begins by establishing a three⁃dimensional finite element model of a pipeline containing buried micro⁃cracks using ABAQUS software,which is then simplified into a two⁃dimensional plane to facilitate crack localization. A total of 1353 different buried micro⁃crack scenarios are simulated by varying the crack positions. Two key feature factors are extracted from the collected ultrasonic guided wave signals： the zero⁃frequency component,which exhibits high sensitivity to the nonlinear effects from the interaction between guided waves and micro⁃cracks,and the damage index,which quantifies the extent of pipeline damage based on the time⁃of⁃flight difference between damaged and intact signals. These factors are fused to create a comprehensive feature matrix,which is then fed into a CNN model for training and testing.The results demonstrate that the zero⁃frequency component is significantly more sensitive to micro⁃cracks than conventional second harmonic components,capturing weak nonlinear effects that are often missed by traditional linear inspection methods. The damage index effectively amplifies the micro⁃crack⁃induced damage information in the ultrasonic signals,enabling precise quantification of pipeline damage. The fusion of these complementary features enhances the CNN model's accuracy in localizing buried micro⁃cracks. In conclusion,this research offers a novel solution for micro⁃crack localization in petrochemical pipelines by combining nonlinear ultrasonic guided wave techniques with advanced machine learning algorithms. The proposed method,validated through extensive simulations,shows high precision in identifying the spatial positions of buried micro⁃cracks,contributing significantly to the safety and reliability of pipeline operations in the petrochemical industry.
This paper introduces an approach to design acoustic metasurfaces. It utilizes multiple nonlinear spring oscillator chains to generate harmonics in the radiated sound field. The metasurface unit is a chain of masses connected by two nonlinear springs exhibiting two resonance frequencies. The fundamental and second harmonic components of the vibration amplitude are solved by the multiple scales method. By setting the higher resonance frequency to be n times the lower frequency，and exciting the system at the lower frequency，energy transfers from the low⁃frequency mode to the high⁃frequency mode. The transfer，induced by nonlinearity，leads to a large vibration amplitude in the high⁃frequency mode. The validity of this method are validated through the consistency between the theoretical and numerical results. Parameter tuning，including adjustments to the quadratic nonlinear coefficient，resonance frequency and excitation frequency，highlights the robustness of this nonlinear system. It provides insights into designing general nonlinear metasurfaces. Furthermore，a nonlinear metasurface model is showcased with stronger high⁃harmonic transmission. Our work lays a solid foundation for realizing harmonics in nonlinear spring oscillators，extending the research scope of acoustic metasurfaces into nonlinear dynamics.
Ammonia (NH3) is a common gas in daily life and industrial production. Because NH3 with a high concentration exerts harm on humans，sensors with high performance are required to detect NH3 with trace concentrations. This work presents a surface acoustic wave sensor for detecting NH3 at room temperature. A reduced graphene oxide (RGO) sensitive layer is grown on a 128° YX⁃LiNbO3 piezoelectric substrate，and platinum (Pt) is deposited on the sensitive layer via vacuum sputtering to enhance the sensitivity of the sensor. The experiment shows that the sensor has a high sensitivity，which exhibits a frequency shift of 8.68 kHz towards 50 ppm NH3 at room temperature. It is observed that the sensor possesses a higher sensitivity when humidity is higher. Additionally，a higher sensitivity of the sensor is obtained in argon than that in air. Moreover，the frequency shift of the sensor is linearly related to NH3 concentration，and the sensor exhibits good repeatability and robustness.
Time series forecasting is a crucial data analysis technique with wide⁃ranging applications in transportation，economics，climate studies，etc. It aids in the rational allocation of resources，major risk decision⁃making，and future planning. Recently，the development of machine learning and deep learning methods has brought significant attention to multivariate time series forecasting. However，existing methods often fail to simultaneously capture the complex dependencies between time points and variables. In this paper，we propose DItrans，a novel time series forecasting model that integrates a transpose embedding method with time series decomposition. Initially，we decompose the time series into trend，periodic，and residual components. Following this，we apply transpose embedding to learn the representations using different encoder structures. The transpose embedding method allows DItrans to better capture correlations between multivariate variables，while the decomposition into trend，periodic，and residual components aids in capturing information from neighboring time points. Additionally，DItrans introduces a new，more flexible encoder structure，enabling the model to capture more complex time series features. We evaluate the performance of the proposed model on three real⁃world datasets. Experimental results demonstrate that DItrans outperforms existing methods in terms of both mean square error (MSE) and mean absolute error (MAE). Specifically，DItrans achieves reductions in MSE ranging from 1.71% to 79.28% and reductions in MAE ranging from 0.72% to 57.52% compared to benchmark algorithms.
In the scenario of rumor detection，effectively identifying rumors in newly emerging events using existing datasets is of paramount importance. This requires rumor detection models to possess robustness and generalizability across various data distributions. However，many existing rumor detection methods are trained using random dataset partitioning. While these models achieve high accuracy with the given partitioning，they fail to align with the original purpose of rumor detection，which is to handle unfolding situations. To enhance the robustness of rumor detection models for real⁃time events，this study proposes a novel data partitioning scheme for benchmarking rumor detection methods. The proposed scheme leverages the temporal occurrence of the events within the dataset and divides the data into training，validation，and test sets based on the event order. To evaluate its effectiveness，we apply this scheme to widely used open-sourced datasets and conduct extensive experiments using six state⁃of⁃the⁃art rumor detection models. The experimental results demonstrate a significant decrease in model accuracy，with a maximum drop of 53% on the datasets we introduced. These results indicate that the proposed dataset partitioning scheme can provide fresh insights for early rumor detection，offering a more realistic and meaningful solution. By considering the temporal dynamics of the events，our approach directs rumor detection models to pay more attention to rumor detection in newly emerging contexts，contributing to the advancement of rumor detection techniques in practical scenarios.
Modern drug discovery faces the challenge of virtual screening of large compound libraries,and the core issue is the improvement of speed and accuracy in molecular docking. AutoDock Vina is one of the most popular molecular docking tools. Our Vina⁃GPU and Vina⁃GPU+ methods achieve up to 50⁃fold and 65.6⁃fold acceleration over AutoDock Vina，respectively，while ensuring docking accuracy. In recent years，large⁃scale pretrained models have achieved great success in natural language processing and computer vision. This paradigm also holds significant potential in addressing the major challenges of virtual screening. Therefore，we propose a new multimodal virtual screening method called Vina⁃GPU GT，which combines Vina⁃GPU+ molecular docking technology with a pretrained Graph Transformer (GT) model to enable fast and accurate virtual screening. This method comprises three consecutive steps：(1) knowledge distillation from a pretrained GT model，which predicts existing molecular properties，to train a smaller SMILES Transformer (ST) model,(2) inference of all molecules in the compound library using the ST model，followed by fine⁃tuning the ST model based on active learning rules，(3) virtual screening using the fine⁃tuned ST model. We conducted virtual screening experiments on three important targets and two compound libraries，comparing our method with two other virtual screening methods. The results showed that Vina⁃GPU GT achieved the best virtual screening performance.
When processing the node features of the event graph，it is difficult to obtain comprehensive event features. Meanwhile，since the event development is dynamic，the network ignores the data changes between short time slices during the processing，making it difficult to capture the global temporal characteristics of the events. To address the above problems，an incident prediction model (Dynamic Enhanced Graph Attention Network，DEGAT) based on dynamic enhanced graph attention network is proposed. This model obtains the comprehensive event features of the historical event graph by building a Gaussian enhanced graph attention network (Enhanced Graph Attention Network，EGAT); we input the initial event vector and EGAT output into the linear layer to obtain the time features，and then input different historical time features into the time coding layer combined with multi⁃head attention mechanism and LSTM to obtain the global time features; finally，the global time feature is input into EGAT，and outputs the prediction results after nonlinear transformation. Experimental results on four social burst datasets show that the proposed model improves 3.88% in accuracy and 4.12% in precision compared with the DynamicGCN method.
To address the problem that the correlation filtering model is easy to degenerate and leads to the decrease of tracking accuracy in the process of UAV (Unmanned Aerial Vehicle) tracking ground moving targets，such as occlusion and illumination change of the target scene，a stable tracking real⁃time tracking algorithm for UAV to ground targets is proposed. Firstly，the appearance fusion model is constructed by using the features of FHOG，CN and grayscale，which improves the adaptability to complex scenes. Then，based on the correlation filter tracking，an adaptive local spatio⁃temporal regularization strategy is designed. Spatial regularization is introduced into the tracker to achieve pixel⁃level filter constraints，and time regularization is designed to optimize the filter update. Secondly，channel reliability fusion is performed on the filter，and an adaptive model update strategy is designed to prevent filter degradation and improve the accuracy of target positioning. Finally，the target recovery module is designed to improve the strength of the tracker and better cope with the complex environment. The experimental results show that the proposed algorithm can better adapt to the tracking task of UAV to the complex scene of ground targets and meet the real⁃time requirements compared with the similar literature algorithms.
Attribute reduction is one of the main research issues in formal concept analysis. However，in real application scenarios，datasets often change as time goes on，in that their attributes may dynamically increase or decrease. Existing methods often require recalculations from scratch and can not fully utilize existing attribute reduction results. The lack of fast updating operation methods leads to low efficiency. Therefore，we use matrix information entropy to explore the granular reduction update mechanism when updating formal context attribute sets. Specifically，we first define the object granular diagonal matrix，and based on it，further introduce the object granular diagonal matrix information entropy，the object granular diagonal matrix conditional entropy，and the measure of the inner and outer importance of the attributes based on DMCE (Diagonal Matrix Conditional Entropy)，and discuss the matrix information entropy reduction method. Secondly，we further explore the dynamic update mechanism of the object's granular diagonal matrix when attributes increase and decrease under the dynamic formal context，and develop the corresponding attribute reduction algorithms based on the matrix information entropy. Finally，experimental verification is performed on six datasets of UCI，and experimental results have shown that the attribute reduction algorithm proposed in this paper has superior runtime compared to other algorithms when facing large⁃scale datasets.
Nitrilase，as a class of green biocatalysts of great value in industrial applications，can efficiently catalyze the hydrolysis of nitrile compounds into carboxylic acids. Despite its wide application，the specific catalytic mechanism of nitrilase remains elusive. Previous studies have revealed that the GLU⁃LYS⁃GLU⁃CYS tetrad in the active center of nitrilase plays a pivotal role in the catalysis，where the CYS residue acts as a nucleophile attacking the nitrile，and the ionization of its thiol group is a key step in the reaction. However，the process of deprotonation of CYS has not been clearly illustrated. This study focuses on the nitrilase from Rhodococcus zopfii (RzNIT) and investigates the protonation state of CYS165 when the substrate has not yet entered the enzyme's active site. Through detailed analysis of possible pathways for CYS165 deprotonation，it is confirmed that CYS165 in RzNIT is in its neutral state in the absence of substrate. This finding lays the foundation for further studies on the catalytic mechanism of RzNIT.
Nanomaterials have good foam stabilizing effect. Through modification and synthesis，their foam performance can be further improved，and the oil recovery of tertiary oil recovery foam flooding can be improved. This article uses graphene oxide (GO) as the matrix，and through surface carboxylation modification，ethylenediamine (EDA) is used as a bridge to graft xanthan gum (XG) with salt resistance and viscosity increasing effects as a new foam stabilizer. The morphology，structure，and functional groups of materials at different modification stages were characterized and analyzed using FT⁃IR，XPS，Raman，and TEM. At the same time，the influence of concentration，salinity and pH on the system was investigated. The results showed that compared to the XG system，the new foam stabilizer had a more significant effect with increasing concentration，and had better salt and pH resistance performance. After aging at 85 °C for 30 days，the stabilizer lost 13%，significantly reducing the loss rate. The interfacial tension value of the system reached the order of 10-2，indicating good interfacial performance. At 85 ℃ and 4 MPa，the laboratory experiment of dual tube displacement was carried out. The results showed that the foam system has achieved plugging in high permeability cores，effectively improved the displacement effect of low permeability cores，and ultimately enhanced the oil recovery by 21.11%.
In this study，a Ni⁃based metal hydroxide⁃organic framework (Ni⁃MHOF) coating was applied for the solid⁃phase microextraction (SPME) of nitrated polycyclic aromatic hydrocarbons (Nitro⁃PAHs) in water. The Ni⁃MHOF was prepared by the hydrothermal method and was coated on a stainless steel wire by physical adhesion using silicone glue. The effects of extraction temperature，extraction time，salt concentration，pH，stirring speed，desorption temperature，desorption time and fiber length on extraction performance were studied through single variable control experiment. The extraction conditions were accordingly optimized and an analysis method for nitro⁃PAHs was established based on the Ni⁃MHOF coated fiber combined with gas chromatography. Ni⁃MHOF coating showed markedly higher extraction efficiency than commercial fibers，especially for Nitro⁃PAHs with four rings， with very high enrichment factors (3447~5238). The SPME method based on Ni⁃MHOF coating for nitro⁃PAHs determination showed wide linear ranges (0.01~8.00 μg·L-1，R2≥0.990)，low detection limits (0.001~0.02 μg·L-1)，and high precision (intra⁃day RSD of 5.8%~12.0%，n=8; inter⁃day RSD of 6.7%~11.8%，n=3; multi⁃fibers RSD of 7.4%~12.3%，n=3). The method exhibited good applicability in real water samples including tap water，lake water，and sewage (recovery of 69.9%~112.6%).
From February 2022 to March 2023，a bird survey was carried out on Nanjing University Xianlin Campus，and infrared cameras were set up on the Tianwen hills inside the campus to investigate the mammals on campus. A total of 129 species belonging to 17 orders and 41 families were investigated，including 19 species of national second⁃class protected animals. Among them，passerine birds are the most，with 74 species of 22 families. A total of 7 species of animals were investigated，belonging to 4 orders and 6 families，including 2 species of national second⁃class protected animals，namely roebuck river deer and raccoon dog. The activity rhythm of Sus scrofa was analyzed，and it was found that the daily activity peak of wild boar was from 5:00 to 7:00，and the secondary peak was from 15:00 to 19:00. By analyzing the data of the whole year，it was found that the morning activity peak of wild boar appeared the earliest in autumn，and the night activity peak appeared the latest in summer.The findings highlight the importance of university campuses in protecting urban biodiversity and provide a reference for promoting citizens to protect natural resources.
Photoresponsive materials can change their physical or chemical properties under light stimulation. Because of their high sensitivity to multiple degrees of freedom of electromagnetic waves and their multi⁃dimensional response capabilities，they have attracted much attention in various application fields related to sensing and actuating. In recent years，light⁃driven systems based on light⁃responsive soft materials have gradually undergone the evolution from simple sensing and stimulated motion to continuous autonomous feedback control. This means that flexible materials gradually have the ability to accept and transform energy on the basis of being stimulated and only undergoing morphological changes，so as to achieve autonomous actuation. A series of soft actuation systems based on flexible light⁃responsive materials demonstrate the ability to identify and accurately track energy sources such as light sources，heat sources，and sound sources that change directions in real time. The difficulty in realizing this kind of "phototropism" and "phototaxis" lies in how to make the light⁃responsive flexible material recognize the direction of the stimulus source and act on it，and stop the action when the direction is aligned，and how to make the direction misaligned when it is subject to external disturbance. It can adjust to the light autonomously，thus reflecting the autonomy similar to that of animals and plants to a certain extent. This article reviews the research progress of photoresponsive materials in the fields of sensing，actuating and adaptive regulation in recent years. It aims to summarize the development process of flexible material actuation from responsive to self⁃control，analyze the physical and chemical mechanisms of energy conversion and transfer in the key processes of the cycle from the three components of sensing，actuating and feedback，and look forward to the development and potential of self⁃controlled light⁃responsive material systems in research fields such as bionic actuators and soft robots.
