Requesting Distant Robotic Action: An Ontology for Naming and Action Identification for Planning on the Mars Exploration Rover Mission
Wales, Roxana C., Shalin, Valerie L., Bass, Deborah S., Journal of the Association for Information Systems
This paper focuses on the development of a naming convention and the use of abbreviated names and a related ontology for science work and distant robotic action that comprise requests for a robotic rover during the NASA Mars Exploration Rover (MER) mission, run by the Jet Propulsion Laboratory (JPL). We demonstrate how abbreviated names and an associated ontology support sharing and identifying information among teams and software tools. An ontology of distant action must take into account a dynamic environment, changing in response to physical events and intentional actions, and reflect the influence of context on the meaning of action. The nascent domain of Martian tele-robotic science, in which specialists request work from a rover moving through a distant landscape, as well as the need to consider the interdisciplinary teams involved in completing that work, required an empirical approach. The formulation of this ontology used ethnographic methods and grounded theory to study human behavior and work practice with software tools.
Key Words: ethnography, grounded theory; domain model, work practice, distributed work, planning technology, knowledge elicitation.
People frequently ask other people to do tasks that appear simple, yet, when analyzed, can be quite complex. A colleague says: "We need four copies of this," and typically, other colleagues produce four copies of the document, because they understand the steps involved in copy-making. The success rate may change if the requests become more complex, e.g. asking for copies of multiple originals or making multiple requests concerning the same original. However, an abbreviated name will identify the task successfully, even though the execution is invariably more intricate than a simple one- sentence request implies.
The success of such communication depends on the speaker and recipient having a set of shared concepts and concept labels. They must agree on the meaning of the relationship between copies and an original document as well as the word that labels this relation. Arriving at an agreement on the types of objects, attributes, and relations in a domain is challenging for a number of reasons. First, cooperating disciplines may define object classes differently. For example Bowker and Star (1999) describe differences in the taxonomy for diseases for early 20th century immigration officers and medical doctors, and even differences in the taxonomies across international immigration agencies. A second challenge lies in the evolution of domain knowledge, which at minimum will add distinctions over time. For instance, the numerous B vitamins started out as a single undifferentiated class until scientific work established specific functions and chemical structures for the various enumerated co-enzymes.
The present paper concerns the relationship between domain models for humans and technology. In particular, we examine the models of geologists and their distant robotic extensions (rovers) operating on the surface of Mars during the 2003-2004 NASA Mars Exploration Rover (MER) mission, run by the Jet Propulsion Laboratory (JPL). Designers of technology recognize the domain model as a critical variable in software development, and the source of variability across programmers (Hadar and Soffer, 2006). Human Factors specialists know that the types of entities that technology incorporates can influence human comprehension of that technology. For example, in modern trajectory planning software for commercial aviation, discrepancies between a pilot's stair-step model of descent including specific ground locations and a programmer's curved model of descent challenges the pilot's ability to understand and use this software (Degani and Weiner, 1997). We add the topic of ontology to the classification scheme for research on human interaction with information systems (Zhang and Li, 2005).
Opportunities for mismatch between technology and human users are increasing as product lifecycle management and ubiquitous computing concerns rise. …