Overview
This work proposes a Generative Adversarial Imitation Learning (GAIL) framework for multi-agent scenarios, integrating an improved social force model to predict pedestrian trajectories. Related work covers path planning and simulation across multiple scenarios.
Scenarios
- PyGAME Room Scenario
- PyGAME Crossing Scenario (Outdoor)
- U3D Outdoor Scenario
- U3D Crossing Scenario #1
- U3D Crossing Scenario #2
PyGAME Room Scenario
Multi-agent path planning and cooperative behavior testing in a confined-room scenario.
PyGAME Crossing Scenario
Convergence and avoidance behaviors at corridor intersections.
U3D Outdoor Scenario
Multi-agent coordination and trajectory prediction in outdoor environments.
U3D Crossing Scenario (Outdoor)
Outdoor corridor traversal and crowd dynamics.
U3D Crossing Scenario
Cross-scenario traversal and behavior prediction based on U3D.