Publication

Task-adaptive robot learning from demonstration with Gaussian process models under replication

Journal Article (2021)

Journal

IEEE Robotics and Automation Letters

Pages

966-973

Volume

6

Number

2

Doc link

https://doi.org/10.1109/LRA.2021.3056367

File

Download the digital copy of the doc pdf document

Abstract

Learning from Demonstration (LfD) is a paradigm that allows robots to learn complex manipulation tasks that can not be easily scripted, but can be demonstrated by a human teacher. One of the challenges of LfD is to enable robots to acquire skills that can be adapted to different scenarios. In this paper, we propose to achieve this by exploiting the variations in the demonstrations to retrieve an adaptive and robust policy, using Gaussian Process (GP) models. Adaptability is enhanced by incorporating task parameters into the model, which encode different specifications within the same task. With our formulation, these parameters can be either real, integer, or categorical. Furthermore, we propose a GP design that exploits the structure of replications, i.e., repeated demonstrations with identical conditions within data. Our method significantly reduces the computational cost of model fitting in complex tasks, where replications are essential to obtain a robust model. We illustrate our approach through several experiments on a handwritten letter demonstration dataset.

Categories

humanoid robots, intelligent robots, learning (artificial intelligence).

Scientific reference

M. Arduengo, A. Colomé, J. Borràs, L. Sentis and C. Torras. Task-adaptive robot learning from demonstration with Gaussian process models under replication. IEEE Robotics and Automation Letters, 6(2): 966-973, 2021.