top of page
CLIPE - Deliverable 3.4 Interaction-Ready Agents Using Reinforcement Learning
Learning Interaction with Characters
Ariel Kwiatkowski, Vicky Kalogeiton, Julien Pettré, Marie-Paule Cani
This survey reviews the main modern DRL algorithms, and how they can be used to tackle the main challenges in character animation, which are skeletal motion control of an individual, physically-based character, and higher-level navigation for a group of agents, seen as moving points on a map.
download full paper:
code:
bottom of page