A Sensor Ontology Literature Review
Roger Eastman, Craig I. Schlenoff, Stephen B. Balakirsky, Tsai Hong Hong · 2013
The purpose of this paper is to review existing sensor and sensor network ontologies to understand whether they can be reused as a basis for a manufacturing perception sensor ontology, or if the existing ontologies hold lessons for the development of a new ontology.We develop an initial set of requirements that should apply to a manufacturing perception sensor ontology.These initial requirements are used in reviewing selected existing sensor ontologies.This paper describes the steps for 1) extending and refining the requirements; 2) proposing hierarchical structures for verifying the purposes of the ontology; and 3) choosing appropriate tools and languages to support such an ontology.Some languages could include OWL (Web Ontology Language) [1] and SensorML (Sensor Markup Language) [2].This work will be proposed as a standard within the IEEE Robotics and Automation Society (RAS) Ontologies for Robotics Automation (ORA) Working Group [3]. Overview of Sensor Ontology EffortNext generation robotic systems for manufacturing must perform highly complex tasks in dynamic environments.To improve return on investment, manufacturing robots and automation must become more flexible and adaptable, and less dependent on blind, repetitive motions in a structured, fixed environment.To become more adaptable, robots need both precise sensing for parts and assemblies, so they can focus on specific tasks in which they must interact with and manipulate objects; and situational awareness, so they can robustly sense their entire environment for long-term planning and short-term safety.Meeting these requirements will need advances in sensing and perception systems that can identify and locate objects, can detect people and obstacles, and, in general, can perceive as many elements of the manufacturing environment as needed for operation.To robustly and accurately perceive many elements of the environment will require a wide range of collaborating smart sensors such as cameras, laser scanners, stereo cameras, and others.In many cases these sensors will need to be integrated into a distributed sensor network that offers extensive coverage of a manufacturing facility by sensors of complementary capabilities.To support the development of these sensors and networks, the National Institute of Standards and Technology (NIST) manufacturing perception sensor ontology effort looks to create an ontology of sensors, sensor networks, sensor capabilities, environmental objects, and environmental conditions so as to better define and anticipate the wide range of perception systems needed.The ontology will include: