A robotool plug-in for roboworld

Abstract

Developing robotic systems is a challenging task due to the inherently underlying complexity. Additionally, the lack of customised techniques and tools means that the current practice of software engineering for robotics is outdated. Therefore, model-driven software engineering, as opposed to simulation- and code-centric approaches, has been advocated for robotics. Considering this, the RoboStar framework, via its associated tool (RoboTool), provides a collection of domain-specific languages tailored for different aspects of the development of robotic systems. One of these languages is RoboWorld: a controlled natural language (CNL) for specifying operational requirements. In this work, we develop a RoboTool plug-in dealing with RoboWorld. This plug-in has a graphical user interface that enables the customisation of project-specific dictionaries. Moreover, it provides surface and structural editors of RoboWorld documents. Integration with underlying natural-language processing techniques and tools is transparent, and, thus, hidden from the end user. From an intermediate representation of RoboWorld documents, which is derived automatically, the plug-in also has automatic support for checking well-formedness conditions, and generating a formal CyPhyCircus semantics. The RoboTool plug-in for RoboWorld was validated considering three case studies: a rescue drone, a ranger robot, and a foraging robot.