Contactless Sleep Quality and Vital Signs Monitoring for Autism Spectrum Disorder Detection

This paper presents a novel approach for contactless monitoring of sleep and respiration patterns using an ultrawideband (UWB) radar system to identify Autism Spectrum Disorder (ASD). The system captures raw radar data to extract key features such as respiration signals, sleep postures, and body movements. Features derived from the processed radar data were classified using a pretrained VGG16 deep learning (DL) model, achieving an overall accuracy of 98%. The proposed method provides a non-invasive and efficient solution for early ASD detection, addressing challenges in conventional monitoring techniques and laying the groundwork for scalable, real-world applications for next generation homes and healthcare centres.