Hardware-Accelerated Non-Contact System for Sleep Disorder Monitoring and Analysis

• Published in: Sensors (Basel, Switzerland) Vol. 25; no. 9; p. 2747
• Main Authors: Sravanthi, Mangali, Gunturi, Sravan Kumar, Chinnaiah, Mangali Chinna, Divya Vani, G., Basha, Mudasar, Janardhan, Narambhatla, Hari Krishna, Dodde, Dubey, Sanjay
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 26.04.2025; MDPI
• Subjects: Algorithms; Analysis; Artificial intelligence; Data entry; Decision trees; Digital integrated circuits; Edge computing; FPGA-based accelerators; Humans; Machine Learning; Medical research; Medicine, Experimental; Monitoring, Physiologic - instrumentation; Monitoring, Physiologic - methods; Movement disorders; Patient monitoring equipment; periodic limb movement disorder (PLMD); random forest classifier; Sensors; Sleep - physiology; Sleep disorders; sleep monitoring; Sleep Wake Disorders - diagnosis; Sleep Wake Disorders - physiopathology; India; periodic limb movement disorder (PLMD); random forest classifier; FPGA-based accelerators; sleep monitoring
• ISSN: 1424-8220; 1424-8220
• DOI: 10.3390/s25092747

This study analyzes human sleep disorders using non-contact approaches. The proposed approach analyzes periodic limb movement disorder (PLMD) under sleep conditions. This was conceptualized as data capture using a non-contact approach with ultrasonic sensors. The model was designed to estimate PLMD and classify it using real-time sleep data and a machine learning-based random forest classifier. Hardware schemes play a vital role in capturing sleep data in real time using ultrasonic sensors. A field-programmable gate array (FPGA)-based accelerator for a random forest classifier was designed to analyze PLMD. This is a novel approach that aids subjects in taking further medications. Verilog HDL was used for PLMD estimation using a Xilinx Vivado 2021.1 simulation and synthesis. The proposed method was validated using a Xilinx Zynq-7000 Zed board XC7Z020-CLG484.