System for the high-level mission control of robotic tasks

The general objective of this Master’s thesis is to develop a modular software design to create, control, and supervise mission control routines in a C++ object-oriented programming environment. State machines and petri networks have often been implemented for the development of these control routines using C++ libraries. The library used in this work was the Boost State Machine Language (SML) Library. This library has been used to develop state machines and create missions to control and manipulate robots, among other uses. It presents interesting features with great applicability to robotics. The SML Boost Library was chosen among others for its speed, intuitiveness, and long-time support, as it is still in development. The library still presents some bugs and issues that are being fixed by its developers. New versions are currently being released including fixes and improvements. The Mission Manager Module was developed with the aim to design missions for high level control of robotics tasks. The module implements tools for mission design, a TCP/IP communication module to control each mission from afar, and a test suit for the missions. Missions can be implemented parallel to each other and share common parts. This mission control software was developed as a branch of a large software project for robot control. One mission has been developed and another one is in development with one purpose: the supervision of a device controlling beam loss and radiation leaks along the Large Hadron Collider (LHC), an experimental facility developed at the Conseil Européen pour la Recherche Nuclearie (CERN). The Mission Manager Module has been successfully implemented. It uses a modular approach to develop high level control missions for robotic purposes. The first mission has been implemented and tested at CERN and the second one is still in development. These missions allow the inspection of devices in a manual and automatic way, respectively. Interesting improvements can be introduced in future developments. Among them, the use of better coding practices and a more systematic approach to the state machines, allowing for more code reusability and strongly improving module performance.