Training toddlers seated on mobile robots to drive indoors amidst obstacles

Abstract

The goal of this research is to train children seated on mobile robots to purposefully and safely drive indoors. Our previous studies show that in about six weeks of training, infants can learn to drive directly to a goal using conventional joysticks. However, they are unable to acquire the advanced skill to avoid obstacles while driving. This limits mobility impaired children from exploring their home environment safely, which in turn impacts their cognitive and social developments in the important early years. In this paper, we describe results where toddlers are trained to drive a robot within an obstacle course. Using algorithms based on artificial potential fields to avoid obstacles, we create force field on the joystick that trains them to navigate while avoiding obstacles. In this `assist-as-needed' approach, if the child steers the mobile robot outside a force tunnel centered around the desired direction, the driver experiences a bias force on the hand. The results suggest that force-feedback joystick results in faster learning than with a conventional joystick.