A model-driving approach to constructing robotic systems

Abstract

Most robotic systems tend to be complex to maintain and reuse because existing frameworks are based mainly on code-driven approaches. This means the software development process is reduced to the implementation of systems using specific programming languages. During the constant evolution, the systems grow in size and in complexity. Even when these approaches address the needs of robotic focused markets, currently used methodologies and toolsets fail to cope with the needs of such complex software development process. The general objective of our work is the definition of a methodological framework supported by a set of tools to deal with the requirements of the robotic software development process. A major challenge is to make the step from code-driven to model-driven in the development of robotic software systems. Separating robotics knowledge from short-cycled implementation technologies is essential to foster reuse and maintenance.