ARTEMIS prosjekt R5COP, Reconfigurable ROS-based Resilient Reasoning Robotic Cooperating Systems

The project goal is to develop reconfigurable hardware and software together with new methods for modular design. This will reduce costs and time needed to adapt robots to changes in the manufacturing industry, and reduce the amount of labor needed to design new robotized solutions in other industries. The project will achieve its goal through five main results: 1. A formalised specification of components and standardization of interfaces. 2. A robust version of ROS, the robot operating system 3. Semi-formal methods of hardware and software 4. Reconfigurable components (sensors and logic) for robust reasoning and action in industrial environments 5. Methods for verification of safety and reliability The results will be evaluated in several demonstrators, for market segments including industrial robotics, service robotics, field robotics, and logistics robotics. SP2 comprises the development of different software components that give solution to specific robotic problems, but at the same time fulfil some of the project's objectives, such as reconfigurability, reusability and resilience. The components and the mentioned properties will have to be demonstrated and evaluated by being integrated in different demonstrators. Some of the components developed here include hardware (embedded platforms, hardware accelerators, sensors), and software, such as perception components or HMI. SP4 is devoted to the implementation of the different demonstrators that will integrate the components and tools developed in SP2 and SP3. There are different demonstrators that implement the use-cases, and that are divided according to markets in the following way: - Industrial robots o Wooden parts manufacture o Flexible assembly o Robot co-worker - Professional service robots o Cooperating cleaning o Outdoor telesecurity - Field robotics o Search and rescue UGV o Farming o RUAV cooperation - Logistics robots o Moving trolley o Order picking o Pick-up and delivery The Norwegian partners are responsible for the demonstrator of the "robot co-worker". The robot platform has been designed and manufactured in collaboration with a robot supplier, in order to have the proper robotic system to fit in the use-case's requirements, and the required flexibility so that it can integrate other components developed in the project. The robot is currently in SINTEF's facilities where all the required hardware and software integrations are being performed in order to have the basic skills (autonomous navigation, localization and so) from which the demonstration can be built upon. The application to which the robot co-worker is oriented is a mobile robot performing inspection operations in an oil and gas platform. The robot is able to perform some inspection tasks in a fully autonomous manner, but also collaborate with operators located outside the platform to support them in realizing more complex inspection and situation assessment operations. Together with the rest of the partners of the consortium, the Norwegian partners are developing a number of components that provide critical skills to the robotic application to fulfil the requirements of the application. Those skills are being developed, integrated and eventually tested and validated through a number of experiments. The list of such skills/functionalities are: - 3D sensors suitable for outdoor environment. - A videostreaming solution that optimizes transmission over different available wireless network. - The localization framework that allows for seamless sensor fusion, and robustness under losses of sensor information. - A system based on stereovision that provides egomotion and pedestrian and cyclist detection and tracking. - A multimodal detection system for humans close to the robot, taking advantage on state-of-the-art techniques (such as deep learning). - A social navigation system, that allows the robot to navigate in a human-aware way, which translates into increased safety for human operators, and improved comfort for operators sharing the same physical space with the robot platform. - A map visualization component for intuitive user interface. - A planner that acts in a high level (task/mission level) able to react and replan under unforeseen circumstances - A supervised teleoperation mode (shared autonomy) that allows for an operator to remotely operate the robot platform, while the robot itself guarantees safety (through embedded collision avoidance) and supports the user in challenging manoeuvres (such as narrow passages). - A simplified web-based GUI for easy programming of the robot's missions and monitoring of the execution. - Different navigation behaviours for the robot platform that adapt to the context conditions of the operational environment, such as safety rules or emergency situations.

European manufacturing industry faces increasing product variances resulting as a consequence of frequent innovation, short product lifecycles, small series production, and shrinking production cycles. At the same time, production cost must be continuousl y reduced. Agile, transformable and re-usable automation and robotics is be a key enabler to manage those trends. However, few robotic components are designed for easy adaptation and reuse. To overcome those shortcomings, R5-COP focuses on agile manufact uring paradigms and specifically on modular robotic systems. Based on existing and newly developed methods for a formal modeling of hardware and software components, R5-COP will support model-based design, engineering, validation, and fast commissioning. Furthermore, using existing interface and middleware standards such as ROS, R5-COP will strongly facilitate integration of components from various suppliers. The proposed modular approach will not only be more flexible than state-of-the-art solutions,bu t will also reduce design, setup, and maintenance costs. Flexible use of robots naturally includes their close cooperation with humans. Therefore, robustness and safety are crucial requirements which will be assured by dedicated verification and validatio n methodologies. The formal specification framework will support component suppliers in efficiently verifying and certifying their modules. R5-COP will help to identify and develop reconfigurable key hardware and software components, and to show the feas ibility and capability of the approach in living labs in manufacturing and service demonstrator environment