index.md

Motion Grammar Parser for JastAdd Examples

A testbed for a potential motion grammar¹ implementation using JastAdd. For the source code, check out the git repository.

Introduction

This is an implementation of motion grammars from Dantam¹ using Reference Attribute Grammars. In this implementation, both the context model and the motion grammar are described using attribute grammars.

Motion Grammars in JastAdd

A JastAdd motion grammar comprises three parts. MotionGrammar contains generic nonterminals present in every motion grammar, a second, use-case-specific grammar has the actual nonterminals of the grammar as well as additonal information to construct a parser, and finally a third context grammar that describes the context within which the motion grammar is parsed.

Use cases

• Very simple Load-Unload grammar taken from the original paper ¹ to show a minimal proof-of-concept.
• A robotic use case sorting dynamically appearing objects into boxes by colour (ACSOS demohttps://2022.acsos.org/details/acsos-2022-posters-and-demos/2/Specifying-Reactive-Robotic-Applications-With-Reference-Attribute-Motion-Grammars).
• A similar robotic use case, but with another input "teaching" where to put the objects and different time

Bibliogaphy

1. Dantam, N., Stilman, M., 2013. The Motion Grammar: Analysis of a Linguistic Method for Robot Control. IEEE Trans. Robot. 29, 704–718. https://doi.org/10.1109/TRO.2013.2239553 ↩ ↩² ↩³