An evaluation of human performance in a risk environment

Abstract

The work described in this thesis follows the research methodology of studying human performance with a machine system in action in the real world. Unfortunately proven methodologies for man-machine systems research do not exist. Nevertheless, by modifying the "Meister Taxonomy" a credible method was achieved for use with the man-machine system model selected for research. The intent of the research was to observe and record the operators' performance whilst they operated a machine system element of the European Space lab during its development and test program, i.e. their performance in terms of attention,recognition, decision making and responses to the warning system-signal stimuli presented by out-of-tolerance machine system conditions. From the observed and evaluated data several aspects of human performance were found to be at odds with prior laboratory research, primarily because such research was fragmented. For example, although attention is the triggering mechanism of the perception process, the process per se depends upon stimuli detection, fields of view and the visual acuity of the operators themselves. By comparison recognition is a more complex process. For example, with a simple machine system, recognition of signal-stimuli by an operator may be satisfied just by spatial location of the signal lamps. Whereas for complex machine systems operators may need to know not only what malfunction triggered the signal stimulus, but also what the cause of the malfunction was. Resolution would therefore depend on the memory storage capabilities of the operators knowledge of the machine system, which introduces subtleties on data storage(retention) and recall. The observed and recorded research data once collated was used to produce operator performance profiles, indicating on these both the operators predicted and evaluated performance, in an attempt to determine which of the two methods is the more valid for this type of research.