Effect of interface type in the VR-based acquisition of pedestrian skills in persons with ASD.

Abstract

Possession of `social' skills is crucial for persons with autism spectrum disorders (ASD) to maintain a certain independence and a better quality of life, and interaction with virtual environments seems an effective learning aid. In a previous study, we reported that in adults with ASD interaction with a virtual environment (a virtual city) is beneficial to the acquisition of pedestrian skills (street crossing and street navigation). Interaction was based on a gesture-based interface (Microsoft Kinect). Here we compare the learning performance when the same virtual environment is operated by a gamepad interface. We used exactly the same training protocol and data analysis than the original study. We found that both interface types are effective in the acquisition of street crossing and city navigation skills. The gamepad interface seems easier to use (thus leading to faster interaction), but gesture-based interfaces are superior in terms of transfer of the learned skills to real road environments (as reported by parents and caregivers).