As global aging continues to accelerate, individuals with disabilities have become important participants in many activities. During emergencies, they are often the biggest victims. Therefore, ensuring the evacuation safety of individuals with disabilities and studying their movement characteristics in crowds are crucial. To overcome the challenge of conducting movement experiments involving pedestrians with real disabilities, a heterogeneous crowd movement experimental platform involving pedestrians with and without simulated disabilities was designed and developed.

The experimental platform comprised a basic building structure, a camera system, equipment to simulate disabilities, and a human motion trajectory extraction and correction system. The basic building structure was constructed with movable boards, and their size and layout could be changed according to various scene and condition requirements. The camera system was primarily composed of several high-resolution cameras (1 280×720 pixels and a frame rate of 25 frames per second, memory cards, and a power supply. The cameras were fixed above the experimental area for the shooting and recording of the entire process. There was equipment for pedestrians with simulated disabilities. Healthy pedestrians wore walking aids or limited movement equipment to simulate disabilities, including crutches, knee pads to limit knee bending, canes, sandbags, wheelchairs, eye masks, and hats. The human motion trajectory extraction and correction system adopted digital image processing technology based on the mean-shift algorithm to extract and track pedestrian motion trajectories. Combined with literature research, controlled experiments, image processing, and the Voronoi diagram method, the experimental scheme of the directional movement of heterogeneous crowds with pedestrians with simulated disabilities in the channel was designed to simulate the real movement process of heterogeneous crowds.

Based on the theoretical analysis, the basic movement rules of heterogeneous crowds were analyzed from the perspectives of human movement trajectory, typical behavior characteristics, speed, density, and flow rate. Some typical behaviors were found in the experiment, such as overtaking and helping behaviors. The speed in the measurement area first dropped rapidly from approximately 1.5 m/s to approximately 0.5 m/s and then maintained a relatively stable state. The density first rose to approximately 1.5 ped/m², fluctuated up and down, and finally dropped rapidly to close to 0. At the same density, a positive correlation was observed between the flow rate and the ratio of pedestrians with helping behavior, especially in the case of higher density. Under the same density, the flow in the channel decreased as the proportion of simulated disabilities increased, with the decline becoming more pronounced at higher densities. Thus, in public places with relatively high crowd densities, providing specific barrier-free access facilities for pedestrians with disabilities and separating the passage of pedestrians with disabilities from healthy pedestrians can effectively improve the efficiency of crowd movement.

The results were consistent with the empirical conclusions and verified the feasibility and rationality of the platform to simulate the movement of heterogeneous crowds with and without disabilities. The experimental platform can provide a reference for experimental research and teaching innovation in traffic safety, safety management, and other fields.