Acoustic source localization(ASL)and sound event detection(SED)are two widely pursued independent research fields.In recent years,in order to achieve a more complete spatial and temporal representation of sound field,...Acoustic source localization(ASL)and sound event detection(SED)are two widely pursued independent research fields.In recent years,in order to achieve a more complete spatial and temporal representation of sound field,sound event localization and detection(SELD)has become a very active research topic.This paper presents a deep learning-based multioverlapping sound event localization and detection algorithm in three-dimensional space.Log-Mel spectrum and generalized cross-correlation spectrum are joined together in channel dimension as input features.These features are classified and regressed in parallel after training by a neural network to obtain sound recognition and localization results respectively.The channel attention mechanism is also introduced in the network to selectively enhance the features containing essential information and suppress the useless features.Finally,a thourough comparison confirms the efficiency and effectiveness of the proposed SELD algorithm.Field experiments show that the proposed algorithm is robust to reverberation and environment and can achieve higher recognition and localization accuracy compared with the baseline method.展开更多
We present a study on the single event transient （SET） induced by a pulsed laser in different silicon-germanium （SiGe） heterojunction bipolar transistors （HBTs） with the structure of local oxidation of silicon ...We present a study on the single event transient （SET） induced by a pulsed laser in different silicon-germanium （SiGe） heterojunction bipolar transistors （HBTs） with the structure of local oxidation of silicon （LOCOS） and deep trench isolation （DTI）. The experimental results are discussed in detail and it is demonstrated that a SiGe HBT with the structure of LOCOS is more sensitive than the DTI SiGe HBT in the SET. Because of the limitation of the DTI structure, the charge collection of diffusion in the DTI SiGe HBT is less than that of the LOCOS SiGe HBT. The SET sensitive area of the LOCOS SiGe HBT is located in the eollector-substrate （C/S） junction, while the sensitive area of the DTI SiGe HBT is located near to the collector electrodes.展开更多
The local seismicity observed by seismic network in siachen-nubra region during January 2010-December 2012 shows that the middle part of the crust (17 - 40 km) is aseismic. This aseismic layer (17 - 40 km) is sandwich...The local seismicity observed by seismic network in siachen-nubra region during January 2010-December 2012 shows that the middle part of the crust (17 - 40 km) is aseismic. This aseismic layer (17 - 40 km) is sandwiched between two seismically active layers and depicts a good spatial correlation with the observations of low resistivity reported from magnetotelluric studies for the same region. The local seismicity shows a trend along the Karakoram fault and clustering of events in Shyok Suture zone and Karakoram shear zone. The moment magnitude of these events lies between 1.3 and 4.3. Most of these events have been originated in upper crust.展开更多
A new method that is applicable to local seismic networks to estimate the azimuth and slowness of teleseismic signals is introduced in the paper. The method is based on the correlation between the arrival times and st...A new method that is applicable to local seismic networks to estimate the azimuth and slowness of teleseismic signals is introduced in the paper. The method is based on the correlation between the arrival times and station positions. The analyzed results indicate that the azimuth and slowness of teleseismic signals can be accurately estimated by the method. Average errors for azimuth and slowness measurements obtained by this method using data of Xian Digital Telemetry Seismic Network are 2.0?and 0.34 s/(?, respectively. The conclusions drawn from this study indicate that this method may be very useful to interpret teleseismic records of local seismic networks.展开更多
The method for rapidly, precisely and non-invasively localizing functional regions of the brain is a problem in neuromedicine research. Cortical electrostimulation is the optimal localization method during brain surge...The method for rapidly, precisely and non-invasively localizing functional regions of the brain is a problem in neuromedicine research. Cortical electrostimulation is the optimal localization method during brain surgery, with a degree of accuracy of approximately 5 mm. However, electrostimulation can damage the cerebral cortex, trigger epilepsy, and extend the operation time. Studies are required to determine whether cortical motor regions can be localized by wavelet analysis from electrocorticograms. In this study, based on wavelet analysis of electrocorticograms, a selection of algorithms for classification of the mu rhythm in the motor regions utilizing experimental data was verified. Results demonstrated that a characteristic quantity of energy ratio in the reconstructed signal was filtered in the d6 （7.81-15.62 Hz） band prior to and following motion events. A characteristic threshold was considered to be 40%. The accuracy of localization detection was 93%. The degree of accuracy was less than 5 mm. The present study avoided the problems of cerebral cortex injury and epilepsy onset, with an operation time of 60 seconds. Therefore, wavelet analysis on electrocorticogram is feasible for localizing cortical motor regions. Furthermore, this localization technique is accurate, safe and rapid.展开更多
基金supported by the National Natural Science Foundation of China(61877067)the Foundation of Science and Technology on Near-Surface Detection Laboratory(TCGZ2019A002,TCGZ2021C003,6142414200511)the Natural Science Basic Research Program of Shaanxi(2021JZ-19)。
文摘Acoustic source localization(ASL)and sound event detection(SED)are two widely pursued independent research fields.In recent years,in order to achieve a more complete spatial and temporal representation of sound field,sound event localization and detection(SELD)has become a very active research topic.This paper presents a deep learning-based multioverlapping sound event localization and detection algorithm in three-dimensional space.Log-Mel spectrum and generalized cross-correlation spectrum are joined together in channel dimension as input features.These features are classified and regressed in parallel after training by a neural network to obtain sound recognition and localization results respectively.The channel attention mechanism is also introduced in the network to selectively enhance the features containing essential information and suppress the useless features.Finally,a thourough comparison confirms the efficiency and effectiveness of the proposed SELD algorithm.Field experiments show that the proposed algorithm is robust to reverberation and environment and can achieve higher recognition and localization accuracy compared with the baseline method.
基金Supported by the National Natural Science Foundation of China under Grant Nos 61274106
文摘We present a study on the single event transient （SET） induced by a pulsed laser in different silicon-germanium （SiGe） heterojunction bipolar transistors （HBTs） with the structure of local oxidation of silicon （LOCOS） and deep trench isolation （DTI）. The experimental results are discussed in detail and it is demonstrated that a SiGe HBT with the structure of LOCOS is more sensitive than the DTI SiGe HBT in the SET. Because of the limitation of the DTI structure, the charge collection of diffusion in the DTI SiGe HBT is less than that of the LOCOS SiGe HBT. The SET sensitive area of the LOCOS SiGe HBT is located in the eollector-substrate （C/S） junction, while the sensitive area of the DTI SiGe HBT is located near to the collector electrodes.
文摘The local seismicity observed by seismic network in siachen-nubra region during January 2010-December 2012 shows that the middle part of the crust (17 - 40 km) is aseismic. This aseismic layer (17 - 40 km) is sandwiched between two seismically active layers and depicts a good spatial correlation with the observations of low resistivity reported from magnetotelluric studies for the same region. The local seismicity shows a trend along the Karakoram fault and clustering of events in Shyok Suture zone and Karakoram shear zone. The moment magnitude of these events lies between 1.3 and 4.3. Most of these events have been originated in upper crust.
基金Foundation of Verification Researches for Arm Control Technology
文摘A new method that is applicable to local seismic networks to estimate the azimuth and slowness of teleseismic signals is introduced in the paper. The method is based on the correlation between the arrival times and station positions. The analyzed results indicate that the azimuth and slowness of teleseismic signals can be accurately estimated by the method. Average errors for azimuth and slowness measurements obtained by this method using data of Xian Digital Telemetry Seismic Network are 2.0?and 0.34 s/(?, respectively. The conclusions drawn from this study indicate that this method may be very useful to interpret teleseismic records of local seismic networks.
文摘The method for rapidly, precisely and non-invasively localizing functional regions of the brain is a problem in neuromedicine research. Cortical electrostimulation is the optimal localization method during brain surgery, with a degree of accuracy of approximately 5 mm. However, electrostimulation can damage the cerebral cortex, trigger epilepsy, and extend the operation time. Studies are required to determine whether cortical motor regions can be localized by wavelet analysis from electrocorticograms. In this study, based on wavelet analysis of electrocorticograms, a selection of algorithms for classification of the mu rhythm in the motor regions utilizing experimental data was verified. Results demonstrated that a characteristic quantity of energy ratio in the reconstructed signal was filtered in the d6 （7.81-15.62 Hz） band prior to and following motion events. A characteristic threshold was considered to be 40%. The accuracy of localization detection was 93%. The degree of accuracy was less than 5 mm. The present study avoided the problems of cerebral cortex injury and epilepsy onset, with an operation time of 60 seconds. Therefore, wavelet analysis on electrocorticogram is feasible for localizing cortical motor regions. Furthermore, this localization technique is accurate, safe and rapid.
