Jigsaw-based Benchmarking for Learning Robotic Manipulation


Xiaobo Liu, Fang Wan, Haoran Sun, Qichen Luo, Wei Zhang, Chaoyang Song: Jigsaw-based Benchmarking for Learning Robotic Manipulation. 2019, (Invited Presentation at ICRA 2019 Workshop on Benchmarks for Robotic Manipulation).

Abstract

Benchmarking provides experimental evidence of the scientific baseline to enhance the progression of fundamental research, which is also applicable to robotics. In this paper, we propose a method to benchmark metrics of robotic manipulation using the jigsaw game to address the spatial-temporal reasoning skills for robot learning. In particular, our method exploits a simple set of jigsaw pieces by designing a structured protocol following a typical, functional implementation of robot manipulation, which is also highly customizable according to a wide range of task specifications. Researchers interested in understanding the performances of their hardware, software, and integration can selectively adopt the proposed protocol to benchmark their research outputs on a comparable scale in the functional, task and system level of details. The purpose is to provide a potential look-up table for learning-based robot manipulation, which is commonly available in other established engineering disciplines, to facilitate the adoption of robotics through calculated, empirical, and systematic experimental evidence.

BibTeX (Download)

@workshop{Liu2019Jigsaw,
title = {Jigsaw-based Benchmarking for Learning Robotic Manipulation},
author = {Xiaobo Liu and Fang Wan and Haoran Sun and Qichen Luo and Wei Zhang and Chaoyang Song},
url = {https://www.ycbbenchmarks.com/ICRA2019_workshop},
year  = {2019},
date = {2019-05-23},
urldate = {2019-05-23},
abstract = {Benchmarking provides experimental evidence of the scientific baseline to enhance the progression of fundamental research, which is also applicable to robotics. In this paper, we propose a method to benchmark metrics of robotic manipulation using the jigsaw game to address the spatial-temporal reasoning skills for robot learning. In particular, our method exploits a simple set of jigsaw pieces by designing a structured protocol following a typical, functional implementation of robot manipulation, which is also highly customizable according to a wide range of task specifications. Researchers interested in understanding the performances of their hardware, software, and integration can selectively adopt the proposed protocol to benchmark their research outputs on a comparable scale in the functional, task and system level of details. The purpose is to provide a potential look-up table for learning-based robot manipulation, which is commonly available in other established engineering disciplines, to facilitate the adoption of robotics through calculated, empirical, and systematic experimental evidence.},
note = {Invited Presentation at ICRA 2019 Workshop on Benchmarks for Robotic Manipulation},
keywords = {ICRA, Workshop},
pubstate = {published},
tppubtype = {workshop}
}