RobOps: Développement et déploiement d’applications robotiques par plateformes IdO

Abstract

This thesis presents the design and implementation of a development and deployment platform for robotic solutions, based on an IoT platform aiming to facilitate the application of DevOps methodologies to the robotics domain. Initially focused on the development of robotic solutions, the project evolved to address frustrations encountered during development and deployment, notably, the frequency of repetitive manual operations and build times. The goal was to create a rapidly operational and easily repeatable development environment and an automated, streamlined pipeline aligned with DevOps best practices to enhance teams’ productivity and experience as well as the quality of the solutions. This project represents a significant initial public access application of these methodologies in the robotics domain, demonstrating their relevance and potential to transform how robotic solutions are developed and deployed. The key results of the project presented in the thesis are : • Reduced propagation times : The deployment method via the IoT platform makes possible the parallelization of an operation that used to be serial by nature during manual deployments. The paradigm shift has resulted in an improvement in the propagation time of new versions in a fleet of machines. The experimentation demonstrated a reduction by more than 22 times in the propagation of the new versions of robotics solutions with only five machines in the fleet, which largely counterbalances the roughly 50-second increase in deployment time, representing about 60%. • Reduced frequency of repetitive manual actions : In the automated version of the project, the experimenter doesn’t need to do any manual actions once a pull request is completed because the CI/CD pipeline takes care of all these procedures. In comparison, a manual deployment includes multiple manual actions that must be repeated for every machine in the fleet. Among these manual actions, we find connecting to the machine (typically via SSH), pulling the latest modifications, maybe even copying the new version to the machine, compiling the solution, and finally swapping the running processes with the new ones. This automation of the deployment process reduces the manual workload and the risk of errors. • Improved process fluidity : By integrating DevOps methodologies (continuous integration, continuous testing, and version control), the project mitigated pain points in development and deployment, making the process more automated, transparent, and efficient. Additionally, containerizing the environments greatly reduced the setup time of the development environment. • Contribution to optimizing robotics development The project aligns with a broader objective : improving the development and deployment experience for teams working on robotic solutions by offering a platform that is tailored, agile, and compliant with DevOps best practices. These results demonstrate that implementing DevOps methodologies has a tangible impact on productivity, error reduction, and user experience in robotics development.

Date	27 Sept 2026
Original language	French
Awarding Institution	École de technologie supérieure
Supervisor	Tony Wong (Supervisor)