Handbook of Simulator-Based Training

Eric Farmer, John van Rooij, Johan Riemersma, Peter Jorna

• Publisher: Routledge
• Year: 2017
• Print ISBN: 9780754611875
• eBook ISBN: 9781351931755
• Topic: Education
• Subtopic: Business General

Advances in simulation technology have enabled an interesting amount of training and instruction to be conducted on training simulators instead of on real systems. However, experiences with the procurement and use of training simulators has not always been as successful, often owing to a lack of knowledge of didactics and of training programme development, and also to inadequate simulator specifications. The Handbook of Simulator-based Training represents the first comprehensive overview of the European state of the art in simulator-based training. It also comprises a well-founded and systematic approach to simulator-based training and the specification of simulator requirements. The multi-disciplinary research project described in this book combines the expertise of specialists in human factors, information systems, system design and engineering from 23 research and industrial organizations from five countries - France, Germany, the Netherlands, Spain, the UK. The authors have synthesized and documented the project results to ensure that this handbook provides not only many valuable guidelines, but more importantly a common frame of reference. It will be a key resource for the many specialists who are concerned with simulator-based training: researchers, engineers, and users; military training institutes and training system development departments; military staff responsible for the procurement of training devices and simulators; the simulator industry; the training research community; and the human factors and ergonomics community.

Section III
Training Media Specification

16 Introduction to Training Media Specification

Introduction

Although this handbook addresses specifically the use of simulators as training media, all available media should be taken into account. In Chapter 10, a distinction was made between CBT, SBT, and RBT. Due to technological developments, the distinctions between different media are blurred. Many capabilities previously available only on high-end simulators are now available on desktop systems for desktop prices. Also, many real systems are equipped with embedded training facilities. It is not easy to keep up with these developments. Every selection rule and media categorisation scheme is at risk of becoming obsolete. Different organisations will have different training needs and constraints, and hence different rules and categorisations may apply. Technological developments cause changes in system concepts and operations and thus changes in training needs. These developments make it all the more important to specify training and training programme requirements accurately. In principle, one might simply require suppliers to provide a training solution. However, it is not easy to formalise and test whether a supplier has succeeded in meeting a particular training need. It is usually not sufficient to specify training or training programme requirements. Additional, more detailed, media requirements are necessary, if only to ensure that training media match the existing infrastructure.

The issue addressed in this section is how to translate training programme requirements into functional simulator requirements. As will be shown, resolving this issue requires knowledge about skills, requisite cues, and their implications for simulation. Before reviewing the available knowledge, the research on which this knowledge is based will be discussed briefly.

Research Issues

In the previous section on TPD, an attempt was made to define skills, and different types of skill were distinguished. Clearly, a skill is an attribute of a person. A cue, on the other hand, is an attribute of the task environment. These concepts are not independent: skill is required to be able to use a cue, and a cue is only a cue for those who have the skill to use it. Each skill that has been specified at a sufficient level of detail can be linked to particular cueing requirements. Here a 'cue' is defined as a modality-specific unit of information. Defining 'cue' as modality-specific renders it somewhat more specific than skill. Also, since they are attributes of the environment, it is possible to define cues partly in physical terms. Although this issue will not be pursued further here, the concept of 'cue' may allow us to define the concept of 'skill', and hence the interface between person and environment, more accurately.

Individuals differ in skill proficiency, and hence cue sensitivity. However, individual differences may also exist with respect to the types of cue that are used ('cue utilisation'). This implies that the mapping between skills and cues will often not be one-to-one.

Another problem with linking skills to cues is that many skill domains are not well researched. What is known about cues has mostly been obtained from rather isolated laboratory studies. Moreover, most applied studies have been conducted in the domain of flight skills; as noted in the introduction to Section II, many have used a transfer-of-training paradigm, and have focused on relatively crude assessments of simulator fidelity. Some have confounded the effects of training and simulation. It is therefore impossible to determine whether the results obtained are due to differences in training approach, to differences between simulator and simulated system, or both.

Applied studies are strongly technology-driven. For example, many earlier studies on motion cueing used technologies that have been superseded by more advanced technologies. Such studies (notably those that reported null results) may have to be replicated (see, for instance, the recent review of research on motion cueing by Burki-Cohen et al, 1998). Most applied studies have used rather subjective assessments (e.g., Cooper-Harper ratings). Although current technologies permit relatively unobtrusive and objective measurements of human and system performance under operational conditions, for a variety of reasons it remains very difficult to gain access to operational personnel and systems and to obtain real performance data. Such data are essential for finding out which cues are used.

Resolving these conceptual, methodological, technological, and organisational problems requires a more integrated research approach, comprising laboratory, simulator, and validation studies. Such an approach requires a more common terminology, and interfaces between training research, research on human perception and performance, and simulation technology. Even if such an approach is initiated, the problem remains that skills change in response to new system concepts and technologies. This naturally has consequences for the types of skill to be trained and the types of cue that are relevant and need to be simulated. For instance, there is a growing need for team, supervisory, and remote control skills.

Fortunately, there are some constraints. One set of constraints can be derived from our extensive knowledge of the capabilities and limitations of the human perceptual system. Another derives from knowledge of the characteristics and limitations of human information processing capabilities. Both types of knowledge are increasingly being used in task and system design as, due to extension of technological capabilities, pressures towards optimal human performance continue to increase. These pressures restrict and delineate further the operational cues that operators use.

Owing to our lack of knowledge, the impact of technological developments, and the specificity of training needs, the need to resolve specific issues during the specification of simulators will remain; in other words, most of the issues addressed in this section have not been definitively resolved. Nevertheless, a systematic approach may help to ensure that the correct issues are addressed. Some ideas for such an approach to the specification of training media will be presented. First, however, existing knowledge will be reviewed.

Skills and Cues

For training to be effective, the cues (features or parameters in the environment) used by the operator to perform his tasks should also be present during training. The type and number of cues that are used depend on the task goal, the task strategy, the skill level of the operator, and the environmental conditions. They may also depend on specific operator knowledge of a particular region and on situational awareness at any given moment (people are notoriously opportunistic in trying to achieve their goals).

Cues can be classified as 'instrument' and 'non-instrument'. An instrument or equipment cue is a parameter of the HMI that the trainee and operator must use as an aid in performing their training activities and tasks. A non-instrument cue is a modality-specific parameter of the physical environment that the trainee or operator uses. Instrument cues may be redundant, complementary, or continuous with respect to non-instrument cues. Instrument cues are relatively well documented in system documentation and manuals, and can be subdivided further into display cues and control cues (e.g., control force feedback). Since it is relatively simple to specify and simulate instrument cues, our main concern in this section is with the specification of non-instrument cues. From the definition given, it should be clear that what counts as a cue depends on the training activity or task to be performed, and, even more specifically, on the trainee or operator who executes it. Another useful distinction is between onset, sustained cues, and stop cues. This distinction refers to the occurrence of events and the maintenance of particular states.

Skill Type and Cueing Realism

For some task categories and skill types, full realism is much easier to achieve than for others. Some of these dependencies and their implications are reviewed below for different skill types.

Perceptual and motor skills Perceptual-motor skills impose the highest demands on simulation (Boer, 1991). Owing to the larger number, variety, and dynamics of the non-instrument cues involved, perceptual-motor skills take longer to learn and the level of realism is more difficult to achieve, if it can be achieved at all. This difficulty is partly due to our as yet incomplete understanding of which non-instrument cues contribute to performance and partly to technological limitations in simulating relevant non-instrument cues. Mismatches between the cues generated by the simulator and those available on the real system may result in skills not transferring completely, or even transferring negatively, to the real system. To counteract these hazards, careful simulator tuning is always required and, if possible, an adequate balance should be maintained between training hours spent on the simulator and the real system.

Procedural skills With respect to training procedural skills, a distinction should be made between learning to recognise the conditions for initiating a particular procedure and learning to execute (and terminate) the procedure. The latter is generally much easier, although retention is relatively poor and frequent recurrency training is required to maintain skills at adequate levels. The difficulty in learning when to initiate particular procedures depends on whether or not procedures are triggered by instrument or non-instrument cues. For many standard operating procedures described in operator manuals, instrument cues are available. However, in some cases non-instrument cues may provide additional or earlier information about whether or not a particular emergency procedure should be executed. Learning the timely recognition of such cues is much more difficult and requires a high level of cueing realism. Non-instrument cues may also be relatively more important in the execution of emergency procedures and in handling degraded system modes.

Cognitive skills The main training focus for cognitive skills is to learn to analyse and integrate different types of information and to use the result in monitoring, planning, and decision making. Most of the information used in exercising cognitive skills is in the form of instrument readings and radio communications. Consequently, a high level of cueing realism is relatively easy to achieve. None the less, circumstances should be replicated sufficiently accurately to induce the appropriate cognitive sets and to learn to deal with the dynamics and complexities of real-life conditions.

Time-sharing and team skills Although the learning of time-sharing and team skills does not directly lead to requirements with respect to realism, in an indirect sense training these skills certainly has implications for the realism required. The amount of transfer will depend upon the extent to which these skills are trained under conditions and task loads that are representative of real-life conditions. A specific requirement of time-sharing skills is that they may demand more attention to be paid to accurate synchronisation of cues. For training particular team skills, e.g., obstacle avoidance, the use of common external spatial references between crew members may require cues to be perceived with specific spatial and temporal accuracy. This has implications for image simulation, e.g., multiple image generators, a larger viewing volume, and separate display systems.

With increased automation, the shift towards procedural and cognitive modes of control will continue. Situational awareness and crew co-ordination will assume a more important role. The specification problem may become more difficult; with the increasing comp...

Table of contents

• Cover
• Half Title
• Title
• Copyright
• Contents
• Preface
• General Introduction
• SECTION I: TRAINING NEEDS ANALYSIS
• SECTION II: TRAINING PROGRAMME DESIGN
• SECTION III: TRAINING MEDIA SPECIFICATION
• SECTION IV: TRAINING EVALUATION
• EPILOGUE
• References
• Glossary
• Index