Cognitive Workload in Fault Diagnosis.
Abstract
Prior research has produced a capability to project with reasonable accuracy the manual activities required to accomplish diagnosis and repair of particular faults within a specified system. When applied to a substantial sample of faults, the technique, termed PROFILE, provides an assessment of the manual workload expected to maintain the system. To obtain usable projections of total repair times, however, requires the estimation of the cognitive portion of fault diagnosis time as well. It is necessary to explore the variations in cognitive times in relation to those characteristics of the diagnostic situation which can be quantified. Three such characteristics are analyzed in detail: 1) the number of tests required to isolate a fault, 2) the total manual time expended in fault diagnosis, and 3) the times of the individual tests in a diagnostic sequence. The most reliable predictor of actual cognitive time for a particular fault was obtained by computing an estimated cognitive time prior to each test based upon the manual time required to perform that test. Since the PROFILE troubleshooting model produces good projections of the particular tests performed to accomplish diagnosis and repair, and their time durations, it can now also compute the estimated cognitive time preceding each test. The sum of the projected cognitive and manual times correspond well with the experimentally observed total diagnosis and repair times. Keywords: Cognition, Troubleshooting, Fault diagnosis, Workload, Maintainability.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 1985
- Accession Number
- ADA165263
Entities
People
- Douglas M. Towne
Organizations
- University of Southern California