RoSta - Robot Standards and Reference Architectures (Jan. 2007 to Feb. 2009) is a Coordination Action (CA) funded under the European Union’s Sixth Framework Programme (FP6). The objective of RoSta is to proactively take the initiative on the definition of formal standards and the establishment of “de facto” standards in the field of robotics, especially advanced service robotics. The project is intended to take the initiative in the formulation of standards in selected key topics which are regarded to have high impact on future service robotics research and development: glossary/ontologies, robot architectures, middleware and benchmarks. These areas include glossary/ontologies, robot architectures, middleware and benchmarks. The topics have been started to form the root of a whole chain of standard defining activities going beyond the specific activities of RoSta. Major RoSta-activities and results can be summarized as follows:
Creation of a glossary/ontology for mobile manipulation and service robots. Comprehensive, widely accepted glossaries form the basis for any standard defining activity. In this context, ontologies (a collection of terms, concepts and their inter-relationships, represented in a machine-usable form) are a means to discuss and formally describe advanced robotics systems in terms of their requirements, functional components, performance, architectures, models and methods. A proof of concept of a robotic ontology scheme and its exemplary use in industrial and service robotics domains demonstrated the benefits of using ontologies in robotics research and developments.
Specification reference architecture principles for mobile manipulation and service robots. The abundance of today’s “home-grown” or “ad-hoc” architectures leads to a divergence rather than to a convergence of robot technology and can be viewed as counterproductive for the scientific progress as much as for the development of new robotic products. Other industries, initially in a similar situation, have established consortia to define reference architectures for complex mechatronic systems (e.g. automotive, aircraft). Due to severe demands on flexibility (in robotics compared to other mechatronic fields), rather than having one standard architecture, the community needs are more in the direction of architectural mechanisms and principles. That should promote efficient engineering and reuse of components, in terms of technical platforms enabling the development of specific architectures (e.g. for mobile manipulation, typically reflecting a certain type of business). A developed action plan suggests a first post-RoSta phase with follow-up work in recently started FP7 projects, and new initiatives thereafter in close collaboration with industrially-driven robotics platform efforts.
Specification of a middleware for mobile manipulation and service robots. Robots are complex systems composed of heterogeneous hardware components with a broad range of communication requirements in addition to varying real-time requirements and distributed processing power over a network of embedded computers. Thus, seamless integration and communication of the various system components require suitable middleware. Existing middleware solutions are regarded too voluminous and inefficient to be used as the standard middleware in mobile manipulation. Results of this activity have been ongoing standard defining activity in “data logging for robotics”, action plan reflected in recently launched FP7 projects on robotics software and a White Paper recommendation: “The use of reuse of HW and SW components in robot development”. The main objective of the paper is to provide insights on flexible and well structured software for robotics, and middleware per se, for a wide class of future robotic systems. Furthermore, many currently existing middleware, control architectures and frameworks have been evaluated on the grounds of specific requirements of advanced robotics.
Benchmarking. Assessing the performance of both a robot component and a system is of strategic importance to facilitate the communication process between research and industry, to document the research progress in robotics and related fields, and to trigger research towards key-functionality performance. In the domain of robotics and neighbouring technologies, numerous documented ad-hoc experimental procedures, benchmarks, and competitions are systematised. Procedures are suggested to develop and use specific benchmarks on different system levels: from evaluating performance of robotic functions via behaviours to assessing full systems in their environments.
These four bodies of activities have been conducted in respective workpackages (WPs) which all followed a similar process starting from an analysis of the state-of-the-art, via requirement analyses, to the formulation of initiatives, action plans and, where appropriate, elaboration of proof-of-concept demonstrations. It can be concluded that RoSta pinpointed central challenges in advanced robotics development as envisaged in the FP6 IST 6th Call on Advanced Robotics. Also through multipliers in the RoSta-consortium to robotics stakeholder groups the project helped initiate sustainable activities in the area of advanced robotics and cognitive systems:
Addressing standardization activities with respect to modular design of robotic systems. Research and standardization of robotic architectures and middlewares has become a prime topic in order to achieve required robot system performance, effectively develop and integrate complex mechatronic systems, and to apply novel engineering methods and tools. Furthermore the emergence of a robotics supply industry as a critical factor for reducing high development costs is based on hardware and software component interoperability and thus on compatibility to standard robot architectures, and middlewares. RoSta helped initiate the concepts of robotics research platforms (as pursued in the newly started FP7-initiatives BRICS and ECHORD) for exploring technologies, methods and tools for an efficient configuration and integration of advanced robot systems for research and industrial applications. Furthermore, standardization activities have started to investigate and specify requirements and interfaces for component interoperability as recently initiated in the context of international standardization bodies such as the ISO TC184, IEEE-RAS and OMG.
Specification of benchmarks. The most critical aspect in benchmarking is their wide-spread adoption among communities. Thus, the discussion and consensus building of the RoSta benchmarking efforts has been taking place in robotics communities such as EURON and EUROP (supported by the CA CARE) as well as in neighbouring benchmarking initiatives (e.g. CA RAWSEEDS). The portfolio of metrics, the process of using existing benchmarks and the development of new benchmarks has been discussed and defined in a sequence of dedicated workshops. Robotics communities are intensely investigating methods of best experimental practice and benchmarking etc. for securing scientific quality and road-mapping of research.
Develop long-term visions and research roadmaps. Of particular interest has been the assessment of maturity levels of robotic key-technologies and the formulation of technology roadmaps in the EUROP SRA. RoSta results regarding basic metrics and maturity levels of robotic key-technologies could be used to quantify a robotics roadmap for the first time.
Study and assessment of international activities and establishment of international cooperation initiatives. The project has been based on intense interaction experts communities (EURON, EUROP, IEEE), reach-out to stakeholder groups and presence in standardisation bodies (ISO, IEEE, OMG) both on a European and international level. Besides using existing formal ties (AIST/Japan and ISRC/Korea) workshops in all areas were usually held within the adjournment of international conferences (ICRA 2007 and 2008, IROS 2007 and 2008, ISR 2008, and ICAR 2009).