GEN AI Enabled Wearable for Early Detection of Sleep Apnea Using NXP I.MX RT1060 and GW1N- LV1LQ144C6/I5 FPGA

View at publisher Summarise with VIRAT
Authors: Karthikeyan A, Balasrinivasan S A, Ashega S, Hamsa Ramyaa P, Harish S S
Cited by: 0
References: 10

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Sleep apnea is a common respiratory disorder that is often left undiagnosed, however, has serious negative impacts on cardiovascular health as well as the quality of sleep in general. This manuscript presents a miniaturized and wearable non-invasive system for sleep apnea detection to realize real-time detection to monitor and intelligently analyze physiological signals continuously. The architecture consists of an NXP i.MX RT1060 microcontroller with a GOWIN GW1N-LV1LQ144C6/I5 field-programmable gate array [FPGA] in order to achieve the parallel acquisition and processing of biosignals. Respiration, snoring and photoplethysmography (PPG) are monitored by AD8233 analog front end, CMM-4030DB-26354-TR MEMS microphone and MAXM86161EFD+T PPG sensor, respectively. Signal conditioning (Switching Kalman filtering, Adaptive filtering Verilog HDL on the FPGA verification with Cadence SimVision in simulations and stable waveform tracking and accurate switching of models. Features extracted have been categorized using a Convolution Neural Network (CNN) which is lightweight and has been implemented on the microcontroller, and it achieved a detection accuracy of 94.6 %. On detecting aberrant respiratory patterns, the system provides realtime haptic feedback via an$8 ~\text{mm} * 2 ~\text{mm}$ERM coin vibration motor as well as Generative AI assisted plain language summaries, along with actionable suggestions, including explanations for apnea episodes, guidance on sleeping positions, alcohol avoidance, and encouragement to consult a physician. Powered with 3.7V 500mAh Li-Po Battery (WLY752530), device works in an energy-efficient manner. Experimental results validate the low latency (approximately 1.2 milliseconds) and very precise results for apnea detection that provide valuable health feedback for actions, which has demonstrated a cost efficient and power efficient algorithmic optimized human health monitoring using signal processing on FPGAs and artificial intelligence embedded in the system.

Fields of Study & Subjects

Non-Invasive Vital Sign Monitoring Obstructive Sleep Apnea Research Embedded Systems and FPGA Design

Publication details

Journal / source
International Conference on Intelligent Systems for Communication, IoT and Security (ICISCoIS)
Pages
1–6
Publication date
8 January 2026
Language
EN