Dynamic Trajectory Forecasting in Crowded Spaces: An in-Depth Review of Online and Adaptive Prediction Methods
As they navigate city streets, autonomous cars will unavoidably collide with people. Research on pedestrian trajectory prediction is crucial for preventing route conflicts with pedestrians. This study summarizes the current public dataset for pedestrian trajectory prediction and compares the performance of various methods. It also examines the pros and cons of depth learning-based trajectory prediction methods. In conclusion, we look forward to the current state of pedestrian trajectory prediction and the difficulties and trends in its growth.
Keywords:
Pedestrian trajectory prediction
Automatic driving
Deep learning
Prediction method
Neural -network
Download & Resources
Article Details
Pages
87 - 102
Views
185
Downloads
130