jcT j jc Obstacle ov & Target Movement ni v & niT i jcT j jc Obstacle ov & Target Movement ni v & niT i (a) t (b) t t ' Fig. 4 Movement velocity vector 3. VALIDITY VERIFICATION OF THE DEM-BASE TRAFFIC MODEL We simulated and experimented with avoiding obstacles and other pedestrians to verify the effectiveness of the constructed traffic model. The avoidance simulation and experiment were done in three types of conditions: (1) obstacle avoidance, (2) overtaking another pedestrian, and (3) avoiding an oncoming pedestrian. The avoidance simulation and experiment are described as follows. 3.1. AVOIDANCE SIMULATION Table 1 shows initial X-Y coordinates of two particles, the direction of the target, and the maximum velocity in each condition. In this simulation, the particle radius was assumed to be 0.2 m, and the time step was assumed to be 1.0×10-4 s. In addition, the maximum velocity used was max v =0.86 m/s, which was actually measured walking speed in this study. Table 1 Parameter of avoidance simulation X-Y coordinate ([m], [m]) Target direction Maximum velocity ( max v ) [m/s] Condition No. No. 1 No. 2 No. 1 No. 2 No. 1 No. 2 (1) (-2,0) (0,0) X 0 0.86 0 (2) (-2,0) (-2,0) X X 0.86 0.43 (3) (-2,0) (2,0) X -X 0.86 0.86 3.2. AVOIDANCE EXPERIMENT An avoidance experiment similar to the avoidance simulations was done. Fig. 5 shows the look of the experiment. By using a motion capture system, pedestrian behavior was measured with four markers attached to the test subject. Fig. 6 shows the attachment position of the markers. Then, the sampling frequency was assumed to be 200 Hz. The default position of test subjects and the target direction were determined as well as the simulation in each condition, and the test subjects were directed to walk in the target direction naturally. As the test subjects headed in the target direction, a walkway 3 m in width was set using colored cones (Fig.5). Fig. 5 Avoidance experiment (Condition No. 1) 234
RkJQdWJsaXNoZXIy MTMzNzEzMQ==