Industrial Robot Ontology
Industrial Robot Ontology (Craig Schlenoff et al., NIST)
Previously, the group has been working on the assumption that the initial work would focus on the Canonical Robot Command Language (CRCL), which is a robot-agnostic messaging language initially developed at the National Institute of Standards and Technology, and would then be expanded to the more general domain of autonomous kit building. This had the advantage of developing a standard for robot interoperability at a low-level, while verifying that this protocol worked by applying it to a simple, but realistic industrial domain.
During their most recent discussions at ICRA2016, it was brought up that the kitting domain is of limited scope. It was feared that while this domain would be useful in proving out our concepts, it may not garner the industrial support that is necessary for the standard to be put into practice. Furthermore, it was not immediately clear that the kitting ontology would be easily extensible to other domains such as welding or full product assembly. It was also pointed out that CRCL could be thought of as an instantiation of a low-level command schema. Due to these limitations, it was suggested that the group refocus their effort on more cross-cutting terms and structures. The suggestions was made that we formulate a standard that addresses the way that hierarchical planners can communicate between levels of a hierarchy. This would include information such as tasks, plans, actions, skills, etc. and would include preconditions and effects of each action.
In this approach, CRCL would become an instance of the lowest level “command” schema, and kitting would become an application that could be used to test the entire command/plan framework.
This concept was proposed to the group as a whole (including those that were not able to attend ICRA) and there was widespread support for this new approach. The group will be writing the Project Authorization Request (PAR), due in September 2016, to become an official IEEE standards group around this concept of developing a standard robotic hierarchical planning architecture.