Trustworthy ROS Software Architecture for Autonomous Drones Missions: From RoboChart Modelling to ROS Implementation

Abstract

Drones are becoming essential tools in emergency situations such as search and rescue, surveillance, and firefighting. These applications make the development of trustworthy software a priority to increase efficiency, cut costs, and reduce risks, in future replacing human personnel in challenging areas. RoboChart, a platform-agnostic tool, brings numerous benefits to the robotics community by providing a language and high-level tool to describe a model via automatic verification and exhaustive testing of the model. However, when it comes to real robotics it is essential to use testing tools suitable for specific robotic software. In the context of the Robotic Operating System (ROS), it imposes its own design constraints, best practices, and communication requirements. This paper aims to apply the RoboChart modelling approach to autonomous firefighting drones, putting forward a low-level ROS software architecture that aligns with RoboChart models, ensuring the trustworthiness of the system during operation.