Comparison of Potential Critical Feature Sets for Simulator-Based Target Identification Training

Abstract : Effective representation of armored vehicles in simulation displays demands a careful evaluation of human perceptual capabilities. This is especially true for computer-generated target displays, which must provide sufficient detail to allow vehicle identification within limitations of computer processing time and display resolution. Even in image generation and display systems not incurring such limitations, the image detail should not exceed human perceptual and cognitive information processing capabilities. Care must be given to vehicle representation to assure that the features represented and emphasized are those most valuable for identifying targets. The current research compared the effectiveness of two different sets of vehicle features for target identification training. Results showed that the two sets of features, in the context of the training in which they were embedded, produced equivalent levels of target identification accuracy, and both produced large gains in performance. Results also revealed that any effects due to range-specific learning of features were very small relative to the improvement produced by training, and were significant when data for one of the programs were analyzed separately.