著者

大佛 俊泰 山田 百合子 金子 弘幸

出版者

日本建築学会

雑誌

日本建築学会計画系論文集 (ISSN:13404210)

巻号頁・発行日

vol.85, no.778, pp.2591-2599, 2020 (Released:2020-12-30)

参考文献数

11

被引用文献数

1

---

In this paper, a model describing pedestrian behavior was constructed, and estimated using the behavior monitoring survey data (behavior data), which was conducted on a university campus square. We analyzed behavioral characteristics and factors of various choice behaviors. Specific results are as follows. (1) First, we conducted a pedestrian behavior monitoring survey on a university campus square using multiple video cameras. In order to convert the pedestrians' position coordinates on the still screen extracted from the video images into position coordinates on a two-dimensional plane, a novel method was developed based on the neural network model, and detailed pedestrian behavior monitoring survey data including attribute information was created. Using this behavior data, we performed a basic analysis on the pedestrian behavior, and clarified the characteristics of pedestrians' destinations and selected routes. (2) Next, we constructed a multinomial logit model that describes the behavior of selecting canteens as a destination, the behavior of selecting a place to stay at the campus square, and the behavior of selecting routes when moving. Each model was estimated using the behavior data, and good description accuracy was confirmed for each model. The estimated parameters of each model showed that the floor area of canteen has a strong effect when selecting a canteen, the positional relationship with the shortest path affects the choice of staying place, and that the resistance of walking distance is affecting route choice. Furthermore, by estimating models for each group size of pedestrians walking together, it was shown that a group with larger number of pedestrians select a canteen with the larger floor area of canteen due to the possibility of available chairs and tables. (3) Finally, we integrated the models and simulated the pedestrian behavior to estimate the pedestrian behavior in the entire university campus square. By comparing the frequency of traffic in each passage based on the behavior data, it was shown that the behavior of pedestrians in the entire square is accurately estimated by the pedestrian behavior simulation. The future development and issues of this model are summarized as follows. Although the model constructed in this paper was able to describe the behavior data of the specific university campus square, it has not been verified yet whether it can be applied to other university campus squares. If the season or weather are different, the behavioral characteristics are expected to be significantly different. It is our future work to study the applicability and expandability of the models by increasing observation data. Furthermore, in this paper, destination points other than the canteens were directly estimated from the selection ratio obtained from the behavior data without modeling. It is necessary to extend the models to a simulation that can describe more detailed pedestrian selection behavior. Finally, the current model does not consider the interaction among pedestrians (such as overtaking, avoiding, and passing). It is necessary to improve the model so that it takes into account the influence of other pedestrians.

著者

金子 弘幸 大佛 俊泰

出版者

日本建築学会

雑誌

日本建築学会環境系論文集 (ISSN:13480685)

巻号頁・発行日

vol.82, no.742, pp.1051-1059, 2017 (Released:2017-12-30)

参考文献数

10

被引用文献数

1

---

The trajectory analysis of pedestrians using the tracking system of laser scanners is an effective approach for understanding the usage of facility spaces. Carefully looking at the pedestrian trajectories observed in an outpatient reception area of a hospital, we can find some specific patterns of patient behavior. For instance, some patients come in the area from the main entrance, line up at the return visit reception machines, and leave to the consulting room area slowly. Also, some coming from the main entrance, pass through the return visit reception machines, and leave the consulting room area quickly. These examples indicate that behavior patterns and attribute information of facility users can be estimated from their trajectory features such as directionality, staying place and walking speed. However, it requires a heavy load to manually extract the features from pedestrian trajectories and classify them into some adequate patterns. Hence, it is highly desirable to achieve this task automatically. In this paper, we proposed a method of pedestrian trajectory classification using the Restricted Boltzmann machine, by which we can automatically find the inherent features of pedestrian trajectories. This method was applied to an outpatient waiting area of a hospital. Comparing manual and automatic classification, we demonstrated the usefulness and sufficient performance of our proposed method in extracting the feature of directionality, staying place and walking speed. The details are as follows: (1) Modeling of pedestrian trajectory The trajectory data were divided into three-layers composed of 1 m square grids, which were consisting of “Front layer”, “Back layer” expressing the directionality, and “Staying layer” expressing staying places. The restricted Boltzmann machine had input units and binary hidden neurons, by which the feature of the trajectory data were generated after sufficient learning. In setting the number of hidden neurons, the 100 × 8 model, which had 100 neurons in the first hidden layer and 8 neurons in the second hidden layer, was applied by comparing the information entropy of the hidden layer. (2) Model validation In the 100 × 8 model, the degree of coincidence between the results by manual classification and automatic classification was examined. The entropy ratio, which is an index for checking the degree of agreement, was 0.6% in the entropy ratio by manual classification, and 10.6% by automatic classification. The results showed that the manual classification and the automatic classification was in good agreement. In addition, the trajectory distribution diagrams were configured for each machine classification, and the feature pattern diagrams were made by 2nd hidden neurons, which automatically found inherent features. These diagrams demonstrated the effectiveness of our proposed method.