Design of a Virtual Task to Understand the Nature of Collaboration Between Autistic and Neurotypical Adults in Workplace Using Multimodal Data

Secure employment is an essential milestone of adulthood due to its association with self-sufficiency, personal achievement, and financial stability. Over the years, autistic population has been greatly affected by high unemployment rates, which is often attributed to their social skills deficit. Existing vocational facilities employ long-term intervention programs that rely on qualitative data to assess collaboration and social skills. However, qualitative data might be prone to error and bias. In addition, long-term qualitative analysis requires the presence of an observer throughout the program, which can result in high operational cost of such training facilities. This paper proposes the design of a collaborative virtual environment that includes LEGO building tasks that are mapped to nine dimensions of collaboration, which can be assessed by collecting and analyzing multimodal data captured during the tasks. A feasibility study with 6 participants was conducted to test the system. The results showed that the system was able to correctly capture multimodal data that corresponded to different dimensions of collaboration. Self-reported surveys reflected that the proposed system fostered a collaborative environment for the participants to exercise their social skills.