Calibrating a Motion Model Based on Reinforcement Learning for Pedestrian Simulation.
Francisco Martinez-GilMiguel LozanoFernando FernándezPublished in: MIG (2012)
Keyphrases
- reinforcement learning
- model free
- motion analysis
- space time
- image sequences
- motion estimation
- reinforcement learning algorithms
- pedestrian dynamics
- real robot
- motion tracking
- simulation model
- motion model
- object detection
- state space
- learning algorithm
- crowded scenes
- moving objects
- optimal policy
- temporal difference
- motion planning
- fully unsupervised
- multi agent
- video sequences
- learning process
- moving camera
- mobile robot
- human motion
- spatial and temporal
- d scene
- camera motion
- feature points