A Color Preference Recognition System for Children with Autism Spectrum Disorder Based on Fully Flexible Triboelectric Sensors and Artificial Intelligence Technology.

Appropriate color intervention can effectively regulate the autonomic arousal level of children with autism, improve their attention, and mitigate their anxiety. In consideration of the issues associated with traditional intervention methods, including cumbersome procedures and the inability to dynamically track the color-selection behavior of children with autism, this paper presents a fully flexible triboelectric sensor based on liquid metal. This sensor utilizes the current direct writing technology to construct a network-like liquid metal electrode layer, which can maintain excellent electrical conductivity under diverse mechanical deformations and temperature variations. Experiments demonstrated that this triboelectric sensor was capable of effectively detecting subtle variations in touch pressure and exhibited high sensitivity, durability, and structural stability in low-voltage monitoring scenarios. Ultimately, a color preference recognition system incorporating liquid metal flexible sensors was established to conduct color selection detection for children with autism under a light-touch operation without imposing a burden. When combined with machine learning, this system was capable of attaining stable and high-precision color selection recognition, with a recognition rate of 98.125%, offering a low-burden and highly effective technical approach for the individualized color preference assessment of children with autism.