네트워크 기반 보행 예측 기법간의 특성에 관한 비교 연구 - 정적 분석과 동적 분석의 관점으로 -

A Comparative Study on the Characteristics of Network-based Prediction tools for Pedestrian Flow - Focused on the viewpoint of Static or Dynamic Analysis -

Despite the difficulty of predicting complex pedestrian flow, recently there have been considerable researches about the quantitative prediction model based on network structure. However, each models have its own focus on one specific feature in network structure, it is hard to find out the difference among those models, especially in terms of the prediction results' closeness to the real pedestrian volume. Therefore, in this study, a comparison study was conducted among the 4 pedestrian prediction models that have been widely researched in our country. In detail, the simulation results of each model have been compared with the real counted pedestrian volume of the selected target area. The result of this study is that the each pedestrian prediction model - mainly divided into static and dynamic, shows lots of differences which might come from the each models' specific focus of the network structure. The dynamic analytical model shows much closer results to the real pedestrian volume, however it needs more time for the pre-survey for the simulation. On the contrary, the static model is simple. Consequently, there should be careful consideration in selecting pedestrian prediction model suits the purpose of the pedestrian prediction experiment.