Part Task Training Methods in Simulated and Realistic Tasks

Abstract

Part task methods are widely used in training programs when the full target task is too complex or impractical due to cost or availability. However, part task training has had mixed success in transfer to the whole task, in comparison to whole task training. For unmanned systems, the goal is to create training that prepares trainees to control an actual vehicle once one is available. The present research compares part and whole task training options in a simulated training environment and a realistic transfer environment. The task is teleoperating a robotic device to detect and identify vehicles in an urban environment. Part task training on the real task yielded a significant benefit in first trial whole task transfer performance and in trials to achieve transfer criteria relative to training on the simulated environment only. While this gain came at a relative cost in overall training time, it produced a significant training effectiveness gain by reducing the training time required on the more expensive robotic system.