Animation generation for object transportation with a rope using deep reinforcement learning

Abstract

This article presents a reinforcement learning‐based approach to generate animation in which two agents use a rope to collaboratively transport a block. The challenge is that the agents need to master several skills, including approaching the block, using the rope to wrap around it, and then moving the block to a predefined goal position. We propose several reward terms to learn the transportation policy and the adjustment policy that govern the skills of the agents. Experiment results showed that the proposed approach was able to generate various animations in different settings, including rope lengths, block sizes, and block shapes. An ablation test revealed the effects of the reward terms. We also investigated factors that affected the performance of the two policies.

Mastering broom‐like tools for object transportation animation using deep reinforcement learning

“…In deep reinforcement learning, agents rely on the reward they receive to evolve and master skills to transport target objects. The agents have various ways to manipulate passive objects, including using their bodies to push, 19 pulling a cart, 8 and using a rope to wrap around an object 9 . The agents not only need to manipulate the target object but also need to learn to navigate the environment and maneuver the target object to avoid collision with obstacles.…”

“…In addition, constraints can be placed on the agents, forcing them to work together. For example, two agents are mounted on a cart 8 or connected by a rope 9 . In these studies, the agents use their bodies to manipulate a target; or the target is attached to the agents and the target is moved while the agents move.…”