Civil Engineering Reference
In-Depth Information
require manual skills, such as positioning, there is more opportunity for
improvement. Positioning typically involves intricate assembly movements for
alignment and orientation of parts.
2. There is not one correct learning rate. For a task such as grinding, Table 15.4 gives two
different learning rates (82% and 98.5%). This illustrates that grinding tasks can be
very different; some of them involve complex movements for which there is a greater
potential for improvement.
15.7
COGNITIVE TASK ANALYSIS AND WORK ANALYSIS
Work is becoming increasingly cognitive, and HFE needs tools to analyze the cognitive
demands and the type of expertise that an operator needs to be successful at work. The
recent developments in cognitive task analysis (CTA) and work analysis provide
interesting avenues for measurement of cognitive demands. These developments are
maybe the most important in HFE during the last ten years. They are, however, still under
development. In this chapter we describe two approaches: the analytical approach and the
situated approach.
1. The analytical approach is used to solve interface problems in complex systems
(Vincente, 1999). This method is based on Rasmussen's (1986) approach to task
description and task decomposition.
2. The situated approach is based on Klein's (1989) concept of recognitionprimed
decision making. This is more suitable for routine tasks, and is based on task
analytical procedures.
15.8
RASMUSSEN'S ANALYTICAL APPROACH
This methodology was originally formulated for analysis of work in complex envi-
ronments, such as nuclear power plants (Rasmussen, 1983, 1986). It has later been
applied to maintenance work and analysis of emergency situations. To understand how an
operator solves problems, the work task is analyzed in terms of abstraction levels and
systems decomposition level. Table 15.5 shows how a power plant can be represented
using five abstraction levels, from purpose to physical form. Ras-mussen (1987) claimed
that the use of five abstraction levels is sufficient to represent even the most complex
analysis, such as that of a nuclear power plant operation.
The abstraction levels can be derived by asking the questions why, what, and how.
Assume that we are working at abstraction level 3, analyzing connected pieces of
equipment. If you ask the question why (Why is this useful?), you will get to abstraction
level 2, and if you ask the question how (How can this be done?), you will get a detailed
description at abstraction level 4. The questions why, what, and how are used to identify
the functionality of any abstraction level. This analytical approach is hence a way of
analyzing a system in terms of the work functions that it generates.
