The Robotic science has been useful in our lives on many ways. Despite its huge potential, the installed industrial robotic applications are not following the fast development of the academic researches. The prototypes, the platforms and the capabilities of the best practices in robotics belong to specific users and this fact increases the cost and decreases the agility of these systems. So, there is a need of automatic software generation for robotic purposes to deliver robotic solutions to more audience. The Model-Driven Engineering promises an automatic software generation as well as the validation of the software, based on models a user creates. Models are mainly expressed in Domain Specific Languages. These languages are constructed to help domain experts to express easier, to validate their models and to make the development process easier and faster. Model-Driven Engineering techniques could reduce the cost and the time of the robotic software development process, making the robotic solutions suitable and agile. ROS 2 is a popular robotic middleware, used to raise the abstraction layers of a robotic software. It is independent of the hardware and it is used worldwide, mainly for academic purposes. Its complexity and the experience it requires from the user is the main reason it is not utilized as it could be in the industry. This work, in the one hand, studies the capabilities of ROS 2 and in the other hand, studies the capabilities the Model-Driven Engineering and their integration. It is a study on how could ROS 2 extend to a Model-Driven Software Development framework. In this work Generos, a ROS 2 system generation software, is introduced. Generos comes with GRS, a Domain Specific Language, used to express models of ROS 2 systems. Generos is able to provide structured ROS 2 system, requiring only robotic skills from its user. It makes the development process faster, easier and safer as it simplifies the process and it validates the models.