Design and Evaluation of a Mobile Robotic Assistant for Emotional Learning in Individuals with ASD: Expert Evaluation Stage

Abstract

This article introduces the design and assessment of a robotic assistant aimed at generating therapeutic interventions for individuals with ASD. A uniform oval-shaped structure was considered for the design, featuring two arms, leg-like wheels, and a facial screen for displaying facial expressions. Six basic emotional states were generated, enriched with facial expressions, colors, sounds, and movements, aiming to emulate human nonverbal language to the greatest extent possible. The construction process was executed using additive manufacturing technology, specifically 3D printing. Once the functional prototype was developed, its appearance, therapeutic usability, generation of emotional states, and social skill development were evaluated through structured surveys on a Likert Scale. The evaluation took place in two stages: (a) expert peer evaluation, involving 5 experts in ASD, with consensus levels determined using Kendall’s Coefficient of Concordance (\(\omega \)), that show a susbtantial agreement (\(\omega =0.709\)) regarding the robot appearance and slight agreement (\(\omega =0.183\)) in mood generation; (b) perception assessment with individuals who work daily with people with ASD, with 36 participants, and survey validation through Cronbach’s Alpha (\(\alpha = 0.94\)), followed by results analysis using descriptive statistics, which indicates that the robot appearance is suitable for the majority of evaluators, but they differ in the robot dimensions. Outcomes highlighted the robotic assistant’s specific characteristics that warrant adjustments before piloting with ASD individuals. This study presents a replicable protocol for the preliminary evaluation of any technological support geared towards therapeutic interventions, preceding experimental processes involving the target audience.