Pedestrian motion simulation method based on visual perception network deep learning

Abstract

The invention discloses a pedestrian motion simulation method based on visual perception network deep learning, which combines deep learning and reinforcement learning to realize pedestrian motion simulation, simulates human vision by using a model, takes RGB images as input and output motion speeds and directions, and can effectively simulate pedestrian flows in different scenes. Compared with atraditional motion model taking position coordinates and the like as input, the scheme takes the visual images as input, is more similar to real pedestrian motion, can effectively simulate pedestrianmotion in two exit evacuation scenes and pedestrian motion in a one-way corridor pedestrian flow scene, and effectively improves the the scalability and robustness of the pedestrian motion model.

Description

technical field [0001] The invention belongs to the field of computer simulation, in particular to a pedestrian motion simulation method based on visual perception network deep learning. Background technique [0002] Pedestrian motion modeling is the main and effective method to simulate and predict pedestrian motion. In the real world, crowd stampede accidents often occur. For example, in subways, campus classrooms and other crowd gathering areas, when fires, earthquakes, terrorist attacks, etc. occur, crowds are very likely to have accidents such as crowding and stampede, causing huge losses of life and property. Therefore, it is important to simulate and predict pedestrian motion. Authorities (i.e., policy makers, evacuation managers, safety planners, researchers) who have prior knowledge of what is likely to occur in the simulated environment can train first responders to respond successfully to actual events as they occur. [0003] Pedestrian motion models are divided...

AI Technical Summary

Problems solved by technology

[0006] However, the traditional pedestrian motion modeling based on reinforcement learning algorithm is insufficient in terms of scalability and robustness. It needs to be combined with deep learning methods to conduct further research on crowd behavior modeling by using deep reinforcement learning algorithms. Robustness to improve

Images