Eric Wai Ming Lee

An Artificial Intelligence-Based Approach for Simulating Pedestrian Movement

This paper proposes a novel approach for simulating pedestrian movement behavior based on artificial intelligence technology. Within this approach, a large volume of microscopic pedestrian movement behavior types were collected and encapsulated into an artificial neural network via network training. The trained network was then fed back into a simulation environment to predict the pedestrian movement. Two simulation experiments were conducted to evaluate the performance of the approach. First, a pedestrian-collision-avoidance test was conducted, and the results showed that virtual pedestrians with learned pedestrian behavior can move reasonably to avoid potential collisions with other pedestrians. In addition, a critical parameter, i.e., defined as “reacting distance” and determined to be 2.5 m, represented the boundary of the collision buffer zone. Second, a pedestrian counterflow in a road-crossing situation was simulated, and the results were compared with the real-life scenario. The comparison revealed that the pedestrian distributions, erratic trajectories, and density-speed fundamental diagram in the simulation are reasonably consistent with the real-life scenario. Furthermore, a quantitative indicator, i.e., the relative distance error, was calculated to evaluate the simulation error of pedestrians trajectories between the simulation and the real-life scenario, the mean of which was calculated to be 0.322. This revealed that the simulation results were acceptable from an engineering perspective, and they also showed that the approach could reproduce the lane-formation phenomenon. We considered the proposed approach to be capable of simulating human-like microscopic pedestrian flow.

Spatial memory enhances the evacuation efficiency of virtual pedestrians under poor visibility condition

Spatial memory is a critical navigation support tool for disoriented evacuees during evacuation under adverse environmental conditions such as dark or smoky conditions. Owing to the complexity of memory, it is challenging to understand the effect of spatial memory on pedestrian evacuation quantitatively. In this study, we propose a simple method to quantitatively represent the evacueeʼs spatial memory about the emergency exit, model the evacuation of pedestrians under the guidance of the spatial memory, and investigate the effect of the evacueeʼs spatial memory on the evacuation from theoretical and physical perspectives. The result shows that (i) a good memory can significantly assist the evacuation of pedestrians under poor visibility conditions, and the evacuation can always succeed when the degree of the memory exceeds a threshold (); (ii) the effect of memory is superior to that of follow-the-crowd under the same environmental conditions; (iii) in the case of multiple exits, the difference in the degree of the memory between evacuees has a significant effect (the greater the difference, the faster the evacuation) for the evacuation under poor visibility conditions. Our study provides a new quantitative insight into the effect of spatial memory on crowd evacuation under poor visibility conditions.

Intelligent route choice model for passengers’ movement in subway stations

Current practice of designing subway stations usually based on relevant design guidebooks and experiences of the designers. Improper station design may lead to bottleneck areas which may reduce the efficiency of the passenger flow. In Hong Kong, microscopic pedestrian movement models have been adopted to predict the pedestrian flow patterns inside subway stations. However, the route choice decisions are required to be pre-defined by the designers. In reality, a passenger should make the decision based on the visual information he/she received. This study collected the actual pedestrian behaviors from subway stations and adopted support vector machine to simulate the decision making on route choice. The results showed that, with 95 % confidence level, the percentage of correct prediction achieved almost 80 %.

Transition from positive to negative on the leadership effect of the biological particles group

This paper reports the existence of the negative effect of leadership on the collective motion of biological particles and shows that the effect of leadership may transit from positive to negative with the change in the sensing capability of the particles. Simulations were conducted in many scenarios using a simple individual-based model. The results showed that leadership can accelerate the collective motion of the biological particles and play a positive role when the sensing capability of the particles is very limited. However, when the sensing capability of the particles becomes large enough, leadership may actually slow the collective motion of the biological particles. This unusual result suggests that leadership may have a negative effect on the collective motion of biological particles. Our finding provides a new insight into how effective leadership can be achieved in a biological particles group.