Kinematic Motion Retargeting via Neural Latent Optimization for Learning Sign Language

TL;DR

This paper introduces a neural latent optimization method for motion retargeting that effectively transfers human sign language motions to robots, leveraging topological graph modeling for improved accuracy and efficiency.

Contribution

It proposes a novel neural latent optimization approach that incorporates topological graph information for better motion retargeting from humans to robots.

Findings

Achieves accurate retargeting of sign language motions to multiple robots.

Demonstrates improved efficiency and convergence over traditional methods.

Validates effectiveness in both simulation and real-world environments.

Abstract

Motion retargeting from a human demonstration to a robot is an effective way to reduce the professional requirements and workload of robot programming, but faces the challenges resulting from the differences between humans and robots. Traditional optimization-based methods are time-consuming and rely heavily on good initialization, while recent studies using feedforward neural networks suffer from poor generalization to unseen motions. Moreover, they neglect the topological information in human skeletons and robot structures. In this paper, we propose a novel neural latent optimization approach to address these problems. Latent optimization utilizes a decoder to establish a mapping between the latent space and the robot motion space. Afterward, the retargeting results that satisfy robot constraints can be obtained by searching for the optimal latent vector. Alongside with latent optimization, neural initialization exploits an encoder to provide a better initialization for faster and better convergence of optimization. Both the human skeleton and the robot structure are modeled as graphs to make better use of topological information. We perform experiments on retargeting Chinese sign language, which involves two arms and two hands, with additional requirements on the relative relationships among joints. Experiments include retargeting various human demonstrations to YuMi, NAO, and Pepper in the simulation environment and to YuMi in the real-world environment. Both efficiency and accuracy of the proposed method are verified.