Goal-Conditioned Variational Autoencoder Trajectory Primitives with Continuous and Discrete Latent Codes
- PDF / 2,279,880 Bytes
- 10 Pages / 595.276 x 790.866 pts Page_size
- 83 Downloads / 194 Views
ORIGINAL RESEARCH
Goal‑Conditioned Variational Autoencoder Trajectory Primitives with Continuous and Discrete Latent Codes Takayuki Osa1,2 · Shuehi Ikemoto1 Received: 13 July 2020 / Accepted: 7 September 2020 / Published online: 16 September 2020 © Springer Nature Singapore Pte Ltd 2020
Abstract Imitation learning is an intuitive approach for teaching motion to robotic systems. Although previous studies have proposed various methods to model demonstrated movement primitives, one of the limitations of existing methods is that the shape of the trajectories is encoded in high dimensional space. The high dimensionality of the trajectory representation can be a bottleneck in the subsequent process such as planning a sequence of primitive motions. We address this problem by learning the latent space of the robot trajectory. If the latent variable of the trajectories can be learned, it can be used to tune the trajectory in an intuitive manner even when the user is not an expert. We propose a framework for modeling demonstrated trajectories with a neural network that learns the low-dimensional latent space. Our neural network structure is built on the variational autoencoder (VAE) with discrete and continuous latent variables. We extend the structure of the existing VAE to obtain the decoder that is conditioned on the goal position of the trajectory for generalization to different goal positions. Although the inference performed by VAE is not accurate, the positioning error at the generalized goal position can be reduced to less than 1 mm by incorporating the projection onto the solution space. To cope with requirement of the massive training data, we use a trajectory augmentation technique inspired by the data augmentation commonly used in the computer vision community. In the proposed framework, the latent variables that encodes the multiple types of trajectories are learned in an unsupervised manner, although existing methods usually require label information to model diverse behaviors. The learned decoder can be used as a motion planner in which the user can specify the goal position and the trajectory types by setting the latent variables. The experimental results show that our neural network can be trained using a limited number of demonstrated trajectories and that the interpretable latent representations can be learned. Keywords Learning from demonstrations · Imitation learning · Deep learning
Introduction Imitation learning (IL) [1] is an approach that can achieve such intuitive motion teaching. In IL, a user demonstrates how a task is performed, e.g. through kinesthetic teaching (Fig. 1), and the system learns how to generalize demonstrated trajectories to different conditions. Previous studies * Takayuki Osa [email protected] Shuehi Ikemoto [email protected] 1
Department of Human Intelligence Systems and Research Center for Neuromorphic AI Hardware, Kyushu Institute of Technology, Fukuoka, Japan
RIKEN Center for Advanced Intelligence Project, Tokyo, Japan
2
have proposed various w
Data Loading...
