<p> 1. Introduction</p><p> In facilities hosting large crowds, such as railway stations, stadiums, and convention centers, visitors in transit have different destinations. Pedestrian flows are therefore complex, unlike single-direction building evacuations. When multiple flows cross, they can easily become stacked because of the sudden increase in density and the difficulty in avoiding other people coming from different directions. Moreover, walking in a disordered crowd is mentally stressful, and even poses the risk of colliding with others. Solving such congestion is essential for flow efficiency, safety, and user comfort.</p><p> This paper proposes a new flow-control system called a Pedestrian-Roundabout, and reports on its effects through the results of two sequential full-scale experiments.</p><p> </p><p> 2. Pedestrian-Roundabout Proposal</p><p> The Pedestrian-Roundabout is an architectural device developed by the authors to optimize the complex pedestrian flows at intersections. In a Pedestrian-Roundabout, a large obstacle several meters in diameter is placed at the center of a pedestrian intersection and, similar to how an automobile roundabout works, pedestrians are required to walk in one direction around it.</p><p> </p><p> 3. Methodology</p><p> To determine the effects of placing such an obstacle on the crowd flow characteristics and detailed walking behavior at the crossing point, we performed two full-scale laboratory experiments: one in 2018 with 48 participants, and another in 2019 with 96 participants. The shape of the obstacle, pedestrian density, and walking direction constraint were all varied in the experimental conditions to determine the effective factors for traffic flow optimization. A short questionnaire evaluating the psychological quality of the walk was also conducted after each trial.</p><p> A very accurate pedestrian trajectory data associated with each pedestrian's ID was obtained using two-dimensional barcode markers and the authors' original image processing programs. Based on these IDs, each trajectory was connected to the other participants' information, including their psychological walk evaluation.</p><p> </p><p> 4. Results</p><p> Diagrams in Fig. 8 indicate the walking trajectories in some representative cases with an emphasis on walking speed reduction. When four flows crossed in a normal right intersection (Fig. 8 4X-FR-VO), the center part became highly congested with pedestrians walking in different directions (Fig. 9). In such a complex crossing situation, pedestrians needed to perform avoidance behavior, namely, speed reduction and detouring, quite often. According to the questionnaire results, the participants perceived the walk as more difficult when they necessitated stronger avoidance.</p><p> By just placing a cylinder shape obstacle with a 4 m diameter in the center of this intersection (Fig. 8 4X-FR-CI), it was observed that the speed reduction was restrained. Furthermore, constraining the walking direction to one-direction (Fig. 8 4X-RA-CI) reinforces this effect, and the trajectories became smoother due to the less detouring around others. The results indicated also that the 4 m diameter obstacle worked more efficiently than a smaller size 2 m obstacle. When the obstacle had edges in the plane, the flow stacked nearby the corners (Fig. 8 4X-FR-SQ₀, 4X-FR-DI₀, 4X-FR-DI₃₀).</p><p> </p><p> 5. Conclusions</p><p> According to the results of our experiments, placing a large obstacle at the center of a complex intersection and letting pedestrians detour around it have positive effects on streamlining the flows. Thus, by introducing the Pedestrian-Roundabout, a reduction in the congestion and a walking difficulty improvement can be expected.</p>