Human trajectory prediction is essential and promising in many related applications. This is challenging due to the uncertainty of human behaviors, which can be influenced not only by himself, but also by the surround...Human trajectory prediction is essential and promising in many related applications. This is challenging due to the uncertainty of human behaviors, which can be influenced not only by himself, but also by the surrounding environment. Recent works based on long-short term memory(LSTM) models have brought tremendous improvements on the task of trajectory prediction. However, most of them focus on the spatial influence of humans but ignore the temporal influence. In this paper, we propose a novel spatial-temporal attention(ST-Attention) model,which studies spatial and temporal affinities jointly. Specifically,we introduce an attention mechanism to extract temporal affinity,learning the importance for historical trajectory information at different time instants. To explore spatial affinity, a deep neural network is employed to measure different importance of the neighbors. Experimental results show that our method achieves competitive performance compared with state-of-the-art methods on publicly available datasets.展开更多
In this work,we study the development,evolution,and migration of turbulent coherent structures in the turbulent boundary layer at Reτ=630 using time-resolved particle image velocimetry(TR-PIV).Multiple techniques,inc...In this work,we study the development,evolution,and migration of turbulent coherent structures in the turbulent boundary layer at Reτ=630 using time-resolved particle image velocimetry(TR-PIV).Multiple techniques,including multi-scale analysis,conditional averaging,cross-correlation,and spatial-temporal topological analysis are applied to extract the evolution principle,migration trajectory,and convection velocity vector of the targeted coherent structures from a Lagrangian perspective.The spanwise vortex structures with larger scale and intensity at a certain wall-normal height y were the main focus of the present study.In the statistical sense,spanwise vortex structures move away from the wall with the shape changing from a bulge to an ellipse,and finally to a circle.Two straight lines emerge from the mean transfer trajectory curve of the spanwise vortex,in which the horizontal one is located at the viscous sublayer(y^(+)<10),the other is a logarithmic straight line existing in the range of 50<y^(+)<120,and the inclination angle of the tangential migration path is fixed at around 12°.The streamwise convection velocity U_(c)of scaled spanwise vortex structures satisfies U_(c)/U_(∞)=0.5-0.6 below y=0.03δ(i.e.,U^(+)_(c)=11-13 undery^(+)=20).In particular,in the region of 50<y^(+)<120,the velocity growth curves of U_(c)and wall-normal convection velocity V_(c)follow the log-law distribution very well,and the slopes are consistent with that of the log-law region of the turbulent boundary layer.Our observations provide microscopic evidences of the logarithmic-linear distribution of the migration trajectory of spanwise vortex structures.展开更多
The volume of trajectory data has become tremendously huge in recent years. How to effectively and efficiently maintain and compute such trajectory data has become a challenging task. In this paper, we propose a traje...The volume of trajectory data has become tremendously huge in recent years. How to effectively and efficiently maintain and compute such trajectory data has become a challenging task. In this paper, we propose a trajectory spatial and temporal compression framework, namely CLEAN. The key of spatial compression is to mine meaningful trajectory frequent patterns on road network. By treating the mined patterns as dictionary items, the long trajectories have the chance to be encoded by shorter paths, thus leading to smaller space cost. And an error-bounded temporal compression is carefully designed on top of the identified spatial patterns for much low space cost. Meanwhile, the patterns are also utilized to improve the performance of two trajectory applications, range query and clustering, without decompression overhead. Extensive experiments on real trajectory datasets validate that CLEAN significantly outperforms existing state-of-art approaches in terms of spatial-temporal compression and trajectory applications.展开更多
Mesoscale eddies exist almost everywhere in the ocean and play important roles in the ocean circulation of the world. These eddies may cause sound spread singular regions and bring great influences to the upwater ship...Mesoscale eddies exist almost everywhere in the ocean and play important roles in the ocean circulation of the world. These eddies may cause sound spread singular regions and bring great influences to the upwater ship and underwater aircraft. Due to the lack of hydrographic survey datasets, study of mesoscale eddies has been greatly restricted. Fortunately, satellite altimeter provided an effective way to study mesoscale eddies. An automatic detection algorithm is introduced to detect mesoscale eddies of specific intensity and spatial/temporal scale based on satellite sea surface height(SSH) data and the algorithm is applied in a strong eddy activity region: the South China Sea and the Northwest Pacific. The algorithm includes four steps. The first step is preprocessing of the SSH image, which includes elimination of error SSH data and interpolation. The second step is to detect suspected mesoscale eddies from preprocessed SSH images by dynamic threshold adjustment and morphological method, and the suspected mesoscale eddy detection includes two procedures: suspected mesoscale eddy core region detection and suspected mesoscale eddy brim extraction. The third step is to pick out mesoscale eddies satisfied with specified criteria from suspected mesoscale eddies. The criteria include three items, that is, intensity criterion, spatial scale, criterion and temporal scale criterion. The last step is algorithm performance analysis and verification. The algorithm has the capability of adaptive parameter adjustment, and can extract mesoscale eddies of interested intensity and spatial/temporal scale. The paper can provide a basis for analyzing space-time characteristics of mesoscale eddy in the South China Sea and the Northwest Pacific.展开更多
基金supported by the National Key Research and Development Program of China(2018AAA0101005,2018AAA0102404)the Program of the Huawei Technologies Co.Ltd.(FA2018111061SOW12)+1 种基金the National Natural Science Foundation of China(61773054)the Youth Research Fund of the State Key Laboratory of Complex Systems Management and Control(20190213)。
文摘Human trajectory prediction is essential and promising in many related applications. This is challenging due to the uncertainty of human behaviors, which can be influenced not only by himself, but also by the surrounding environment. Recent works based on long-short term memory(LSTM) models have brought tremendous improvements on the task of trajectory prediction. However, most of them focus on the spatial influence of humans but ignore the temporal influence. In this paper, we propose a novel spatial-temporal attention(ST-Attention) model,which studies spatial and temporal affinities jointly. Specifically,we introduce an attention mechanism to extract temporal affinity,learning the importance for historical trajectory information at different time instants. To explore spatial affinity, a deep neural network is employed to measure different importance of the neighbors. Experimental results show that our method achieves competitive performance compared with state-of-the-art methods on publicly available datasets.
基金the Young Scientists Fund of the National Natural Science Foundation of China(Grant No.11802195)the National Natural Science Foundation of China(Grant Nos.12172242,and 11972251)+2 种基金the Key Program of the National Natural Science Foundation of China(Grant No.11732010)Sino-German International Cooperation Project supported by Sino-German Science Center(GZ1575)the Natural Science Foundation for Young Scientists of Shanxi Province,China(Grant No.201801D221027).
文摘In this work,we study the development,evolution,and migration of turbulent coherent structures in the turbulent boundary layer at Reτ=630 using time-resolved particle image velocimetry(TR-PIV).Multiple techniques,including multi-scale analysis,conditional averaging,cross-correlation,and spatial-temporal topological analysis are applied to extract the evolution principle,migration trajectory,and convection velocity vector of the targeted coherent structures from a Lagrangian perspective.The spanwise vortex structures with larger scale and intensity at a certain wall-normal height y were the main focus of the present study.In the statistical sense,spanwise vortex structures move away from the wall with the shape changing from a bulge to an ellipse,and finally to a circle.Two straight lines emerge from the mean transfer trajectory curve of the spanwise vortex,in which the horizontal one is located at the viscous sublayer(y^(+)<10),the other is a logarithmic straight line existing in the range of 50<y^(+)<120,and the inclination angle of the tangential migration path is fixed at around 12°.The streamwise convection velocity U_(c)of scaled spanwise vortex structures satisfies U_(c)/U_(∞)=0.5-0.6 below y=0.03δ(i.e.,U^(+)_(c)=11-13 undery^(+)=20).In particular,in the region of 50<y^(+)<120,the velocity growth curves of U_(c)and wall-normal convection velocity V_(c)follow the log-law distribution very well,and the slopes are consistent with that of the log-law region of the turbulent boundary layer.Our observations provide microscopic evidences of the logarithmic-linear distribution of the migration trajectory of spanwise vortex structures.
基金National Natural Science Foundation of China (Grant No. 61772371,No. 61972286)
文摘The volume of trajectory data has become tremendously huge in recent years. How to effectively and efficiently maintain and compute such trajectory data has become a challenging task. In this paper, we propose a trajectory spatial and temporal compression framework, namely CLEAN. The key of spatial compression is to mine meaningful trajectory frequent patterns on road network. By treating the mined patterns as dictionary items, the long trajectories have the chance to be encoded by shorter paths, thus leading to smaller space cost. And an error-bounded temporal compression is carefully designed on top of the identified spatial patterns for much low space cost. Meanwhile, the patterns are also utilized to improve the performance of two trajectory applications, range query and clustering, without decompression overhead. Extensive experiments on real trajectory datasets validate that CLEAN significantly outperforms existing state-of-art approaches in terms of spatial-temporal compression and trajectory applications.
文摘Mesoscale eddies exist almost everywhere in the ocean and play important roles in the ocean circulation of the world. These eddies may cause sound spread singular regions and bring great influences to the upwater ship and underwater aircraft. Due to the lack of hydrographic survey datasets, study of mesoscale eddies has been greatly restricted. Fortunately, satellite altimeter provided an effective way to study mesoscale eddies. An automatic detection algorithm is introduced to detect mesoscale eddies of specific intensity and spatial/temporal scale based on satellite sea surface height(SSH) data and the algorithm is applied in a strong eddy activity region: the South China Sea and the Northwest Pacific. The algorithm includes four steps. The first step is preprocessing of the SSH image, which includes elimination of error SSH data and interpolation. The second step is to detect suspected mesoscale eddies from preprocessed SSH images by dynamic threshold adjustment and morphological method, and the suspected mesoscale eddy detection includes two procedures: suspected mesoscale eddy core region detection and suspected mesoscale eddy brim extraction. The third step is to pick out mesoscale eddies satisfied with specified criteria from suspected mesoscale eddies. The criteria include three items, that is, intensity criterion, spatial scale, criterion and temporal scale criterion. The last step is algorithm performance analysis and verification. The algorithm has the capability of adaptive parameter adjustment, and can extract mesoscale eddies of interested intensity and spatial/temporal scale. The paper can provide a basis for analyzing space-time characteristics of mesoscale eddy in the South China Sea and the Northwest Pacific.
