Dynamic Movement Primitive based Motion Retargeting for Dual-Arm Sign Language Motions

Liang, Yuwei; Li, Weijie; Wang, Yue; Xiong, Rong; Mao, Yuan; Zhang, Jiafan

doi:10.48550/arxiv.2011.03914

Cited by 2 publications

(4 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Comparative Study: We compare against an optimization based method DMPMR [13], and a feedforward neural network. The feedforward neural network, donated as NN, is composed of four linear layers with channels 128, 256, 128, 14, and uses DMPMR's output as ground truth for training.…”

Section: Resultsmentioning

confidence: 99%

“…By optimizing on a pre-defined objective function, these approaches minimize the gap between the solution and the demonstration during the iterative process. Liang et al proposed a motion retargeting method that leverages graph optimization and Dynamic Movement Primitives [13]. They employed DMPs in a leader-follower manner to parameterize the original trajectories and adopt a three-step optimization procedure to find the desired robot motions.…”

Section: B Optimization-based Motion Retargetingmentioning

confidence: 99%

“…It is a basic problem of learning from demonstration and has recently received great attention since it enables service robots to behave more like humans. Nowadays, it has been applied to humanoid robots in entertainment parks [10] and sign language robots for communicating with hearing-impaired individuals [13].…”

Section: Introductionmentioning

confidence: 99%

“…Yue Wang is the corresponding author wangyue@iipc.zju.edu.cn. ability of the robot [13][14] [6]. Nonetheless, optimizationbased methods have to take extensive time for optimizing each motion and have no guarantee to satisfy robot constraints.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Kinematic Motion Retargeting via Neural Latent Optimization for Learning Sign Language

Zhang¹,

Li²,

Liu³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Motion retargeting from human to robot remains a very challenging task due to variations in the structure of humans and robots. Most traditional optimization-based algorithms solve this problem by minimizing an objective function, which is usually time-consuming and heavily dependent on good initialization. In contrast, methods with feedforward neural networks can learn prior knowledge from training data and quickly infer the results, but these methods also suffer from the generalization problem on unseen actions, leading to some infeasible results. In this paper, we propose a novel neural optimization approach taking advantages of both kinds of methods. A graph-based neural network is utilized to establish a mapping between the latent space and the robot motion space. Afterward, the retargeting results can be obtained by searching for the optimal vector in this latent space. In addition, a deep encoder also provides a promising initialization for better and faster convergence. We perform experiments on retargeting Chinese sign language to three different kinds of robots in the simulation environment, including ABB's YuMi dual-arm collaborative robot, NAO and Pepper. A real-world experiment is also conducted on the Yumi robot. Experimental results show that our method can retarget motion from human to robot with both efficiency and accuracy.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: B Optimization-based Motion Retargetingmentioning

confidence: 99%