Detecting avoidance behaviors between moving object trajectories

详细信息    查看全文

• 作者：Francesco Lettich^a ; ^{lettich@dais.unive.it" class="auth_mail" title="E-mail the corresponding author} ; Luis Otavio Alvares^b ; ^{luis.alvares@ufsc.br" class="auth_mail" title="E-mail the corresponding author} ; Vania Bogorny^b ; ^{vania.bogorny@ufsc.br" class="auth_mail" title="E-mail the corresponding author} ; Salvatore Orlando^a ; ^{orlando@unive.it" class="auth_mail" title="E-mail the corresponding author} ; Alessandra Raffaetà ; ^a ; ^{raffaeta@unive.it" class="auth_mail" title="E-mail the corresponding author} ; Claudio Silvestri^a ; ^{silvestri@unive.it" class="auth_mail" title="E-mail the corresponding author}
• 关键词：Trajectory avoidance detection ; Spatio-temporal data analysis ; Trajectory data mining ; Moving object behavior analysis
• 刊名：Data & Knowledge Engineering
• 出版年：2016
• 出版时间：March 2016
• 年：2016
• 卷：102
• 期：Complete
• 页码：22-41
• 全文大小：2213 K

文摘

Several algorithms have been proposed in the last few years for mining different mobility patterns from trajectories, such as flocks, chasing, meeting, and convergence. An interesting behavior that has not been much explored in trajectory pattern mining is avoidance. In this paper we define the avoidance behavior between moving object trajectories, providing a set of theoretical definitions to precisely describe various kinds of avoidance, and propose an effective algorithm for detecting avoidances. The proposed method is quantitatively evaluated on a real-world dataset, and correctly detects with high precision the quasi totality of the trajectory pairs that exhibit avoidance behaviors (F-measure up to 95%).