A novel haptic training method through skill decomposition

Skill training with haptic guidance is more efficient than traditional ways which are visual and/or auditory guidance only. There have been several haptic training methods presented in previous studies. But how to exert effective guidance force is still a difficulty. This paper proposed a novel training method through skill decomposition. We decompose the task skill into several basic skill components. During training, the novices are controlling one component and are guided by expert in the other ones. Skill components are trained individually in this way. When a novice without cooperation with expert can finish the task very well just like the expert does, he/she acquires a skill component. We set up a 2-DOF control training experiment to compare the novel haptic training method with visual only training and traditional force guidance training. Skill for the training task is decomposed into controlling each DOF. The preliminary experimental result shows that the novices´ short-term skill improvement using the novel haptic training method is better than that using the virtual fixture guidance training and visual training.