Towards a Real-Time Warning System for Detecting Inaccuracies in Photoplethysmography-Based Heart Rate Measurements in Wearable Devices

2025-08-27

AI Summary

PPG-based heart rate (HR) monitoring in wearable devices is highly susceptible to motion artifacts and other confounding factors, leading to degraded accuracy that users cannot readily detect. To address this, we propose the first real-time reliability classification method for PPG HR measurements that relies solely on HR time-series signals—without requiring raw PPG waveforms or auxiliary sensors. Our approach employs a lightweight deep learning model trained and validated on real-world data collected from Polar and Garmin wearables. It enables online identification of inaccurate HR readings and provides interpretable feedback to users. The system achieves over 80% detection accuracy, significantly enhancing user trust in health data and supporting more robust human–machine collaborative decision-making. By offering high practicality and low computational overhead, our method establishes a novel paradigm for reliability assurance in wearable health monitoring.

Technology Category

• Machine Learning: Hardware-aware ML
• Humans and AI: Human-in-the-loop Machine Learning
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Web Mining and Content Analysis: Web data provenance, reliability, and authenticity
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Wearable devices with photoplethysmography (PPG) sensors are widely used to monitor heart rate (HR), yet often suffer from accuracy issues. However, users typically do not receive an indication of potential measurement errors. We present a real-time warning system that detects and communicates inaccuracies in PPG-derived HR, aiming to enhance transparency and trust. Using data from Polar and Garmin devices, we trained a deep learning model to classify HR accuracy using only the derived HR signal. The system detected over 80% of inaccurate readings. By providing interpretable, real-time feedback directly to users, our work contributes to HCI by promoting user awareness, informed decision-making, and trust in wearable health technology.

Problem

• Detecting inaccuracies in wearable PPG heart rate measurements
• Providing real-time warning system for measurement errors
• Enhancing user trust and awareness in health technology

Innovation

• Deep learning model classifies HR accuracy
• Real-time warning system detects inaccuracies
• Uses only derived HR signal for classification