One important advantage of simulators for training purposes is that they can easily be equipped with additional systems that automatically measure task performance. Driver training with utilization of these systems may provide two major advantages above more usual training on a driving simulator: the objectivity of performance evaluation by the instructors and the quality of behavioral feedback to the student may be increased. The present paper is the result of a study aimed at improvement of the Performance and Marking (PAM) system used to measure driving performance in the training of tracked-vehicle drivers of the Netherlands Royal Army. The paper outlines the shortcomings of the PAM system and presents a design for a more user friendly Performance Evaluation and Feedback (PEF) system. Seven principles that may be crucial for a successful development of performance evaluation and feedback systems for training simulators are: 1) task analysis, 2) the validity of the simulator for different subtasks, 3) the relevance of subtasks and of performance measures, 4) the evaluation difficulty of subtasks, 5) the manner of metric construction, 6) the comprehensibility, 7) the ergonomics of data-presentation. In the design of a PEF system these principles should be systematically applied. For the Dutch driving simulator this was done accordingly.
Abstract