(1) Directional: the relations or interactions among purchase OSI-420 the pedestrians have direction, as pedestrians are only influenced by the front pedestrians and would not observe the behavior of pedestrians behind them in most cases. (2) Complex: nodes are complex, as pedestrian itself is a complex individual, whose behavior is influenced by personal factors, other pedestrian’s behavior, and traffic environment. And the links are also complex, referring to the complexity, variability, and randomness of pedestrian behavior, as the relationship between the pedestrians is represented

by their behaviors. (3) Increasing: under the impact of conformity psychology, pedestrian’s violation behavior will increase constantly. The pedestrian network’s nodes will grow dynamically. 3.3. Data Collection Signalized intersections with large pedestrian volume and many illegal pedestrians are selected as the study sites in this paper. Pedestrian crossing behavior is studied through the data collected

by a direct observation of pedestrian activities using two video cameras set up beside the crosswalks. Cameras are placed in relatively concealed locations, such as the nearby billboards and street trees, so that the presence of the camera would not affect the pedestrian’s normal crossing behavior. In spite of pedestrian’s crossing behavior such as violating or not, cameras also could film pedestrian’s microscopic moving activities such as head movement, turning aside, looking, and saccade. After the field study, researchers from Nanjing University of Science and Technology record the information obtained from the video. The detailed procedures to determine the behavior relationship between different pedestrians are presented as follows. Firstly, the pedestrian interaction region from the video is judged. It is shown in the literature [16] that observation range of the pedestrian is oval. And this result can roughly determine the pedestrians’ interaction region. Then, observe the behavior of pedestrians who are in the interaction region to see whether there

are some direct motion interactions between the pedestrians. If the pedestrian conducts some motions such as head movement, turning aside, looking, saccade, and AV-951 talking, the pedestrian is seen to have interactions with other pedestrians. After that, researchers can record the useful data and take notes on the pedestrian’s behavior and related information, such as signal cycle, signal time, pedestrian gender, and influencing pedestrian number. The survey is carried out in the morning and evening peak hours. Video camera is placed at each crosswalk to record the expression and action of the pedestrian clearly. According to the basic method of the behavioral effects of relationship determination, the relationships between the pedestrians in different red light stage, different genders are recorded. Table 1 shows the data record sheet for pedestrian interactions. Table 1 Survey data record table. 3.