Related Experiment Video
Updated: Jul 9, 2026
16:14
Trajectory Data Analyses for Pedestrian Space-time Activity Study
Published on: February 25, 2013
13.5K
Analysis of pedestrian second crossing behavior based on physics-informed neural networks
Yongqing Guo1, Hai Zou1, Fulu Wei2
1School of Transportation and Vehicle Engineering, Shandong University of Technology, Zibo, 255000, China.
Scientific Reports
|September 11, 2024
Summary
Physics-Informed Neural Networks (PINNs) model pedestrian flow at two-stage crossings. This advanced AI accurately predicts pedestrian fluid dynamics, enhancing safety and mobility in busy urban intersections.
Area of Science:
- Civil Engineering
- Transportation Engineering
- Artificial Intelligence
Background:
- Two-stage pedestrian crossings are prevalent at busy signalized intersections to manage high traffic volumes and long crosswalks.
- Effective pedestrian management is crucial for safety and operational efficiency in urban environments.
- Understanding pedestrian behavior, particularly during complex bidirectional interactions, is essential for facility design.
Purpose of the Study:
- To develop a novel Physics-Informed Neural Network (PINN) model for analyzing pedestrian flow at two-stage crossings.
- To incorporate fluid dynamics principles into the model to predict pedestrian characteristics like speed, density, and acceleration.
- To evaluate the model's performance against traditional deep learning methods.
Main Methods:
- A Physics-Informed Neural Network (PINN) model was developed, integrating fluid dynamics equations.
- The model was trained and tested using data from pedestrian crossings.
- Performance was assessed by comparing predicted pedestrian fluid properties (speed, density, acceleration, Reynolds number) against established metrics, focusing on mean squared error.
Main Results:
- The PINN model demonstrated superior performance compared to traditional deep learning methods in calculating and predicting pedestrian fluid properties.
- Achieved a highly accurate mean squared error as low as 10-8.
- The model effectively captured dynamic pedestrian flow characteristics and provided insights into behavioral impacts.
Conclusions:
- PINNs offer a powerful and accurate approach for modeling pedestrian flow dynamics at complex crossings.
- The findings support the design of enhanced pedestrian facilities and optimized signal timing for improved safety and mobility.
- This research contributes to intelligent transportation systems by enabling autonomous vehicles to better predict pedestrian intentions.
