Automation always plays an important role in industry. Today, it is a basic need for industry. To develop faster manufacturing or delivery, automation is an important need. Robots always play the main role for automation in the industry. Robots are mainly designed for specific task. But, the main problem is robots are too expensive for one task. Thats why, it is almost impossible to use robots for small industries. Therefore, we are aiming to develop a pipeline to design a multitasking robot, especially for different kinds of packaging tasks. Typical text-based instruction sheets are the main source of these automation robots, that means robots can pack different types of shapes using typical text-based packaging instructions. In robotics, learning by demonstration in robotics, could benefit from large body movement dataset extracted from textual instructions. The interpretations of instructions for the automatic generation of the corresponding movements thereof are difficult tasks. We examine methods for converting textual surface structures into the semantic representations and explore tools for analysis and automated simulation of activities in industrial and household settings. In our first step, we try to develop a pipeline from textual instructions to virtual actions that includes traditional language processing technologies as well as human computation approaches. Using the resulting virtual actions, we will train robots through imitation learning or learning by demonstration for multitasking packaging robots.
[1]
Minhua Ma,et al.
Automatic Conversion of Natural Language to 3D Animation
,
2006
.
[2]
John B. Lowe,et al.
The Berkeley FrameNet Project
,
1998,
ACL.
[3]
Jerry B. Weinberg,et al.
A Multidisciplinary Model for Using Robotics in Engineering Education
,
2001
.
[4]
Angel X. Chang,et al.
Semantic Parsing for Text to 3D Scene Generation
,
2014,
ACL 2014.
[5]
INVESTIGATION INTO A LOW COST ROBOTIC VISION SYSTEM
,
2015
.
[6]
Christopher D. Manning,et al.
Generating Typed Dependency Parses from Phrase Structure Parses
,
2006,
LREC.
[7]
Michael Beetz,et al.
Robots, Pancakes, and Computer Games: Designing Serious Games for Robot Imitation Learning
,
2015,
CHI.