A group member search method based on similarity analysis of mobile patterns

Abstract

Searching for the group members that participate in a joint action based on collecting and analyzing the position information of moving objects could provide an important support for the decision makers in the fields such as the investigation of criminal cases, group emergency disposal, a large-scale military action etc. Following the pattern analysis of mobile objects based on their position information, this paper proposes a group member search method for mobile objects based on pattern similarity, which identifies group members by utilizing the correlation information of the mobility attributes such as trajectory, speed, direction and acceleration. Compared with group clustering or classification based on trajectory distance, the overlapping interference can be effectively avoided between members in different groups, and this method fits the objective law of group movement better. Combining with the application background, the experimental analysis confirms that the proposed method can quickly and accurately find the group members.