BCG-VARS: BallistoCardioGraphy vital algorithms for real-time systems

• Published in: Biomedical signal processing and control Vol. 87; p. 105526
• Main Authors: Thirion, Adrien, Combes, Nicolas, Mulliez, Blaise, Tap, Hélène
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2024; Elsevier
• Subjects: BallistoCardioGraphy (BCG); Bioinformatics; Computer Science; Heart rate variability (HRV); Real-time and embedded system; Respiratory rate variability (RRV); Vital parameters measurement system; Vital parameters measurement system; Heart rate variability (HRV); Respiratory rate variability (RRV); BallistoCardioGraphy (BCG); Real-time and embedded system
• ISSN: 1746-8094; 1746-8108
• DOI: 10.1016/j.bspc.2023.105526

Disorders with a transient signature (e.g., obstructive sleep apnea, drowsy driving, or atrial fibrillation) are difficult to predict. In addition, the aging population is moving towards dependency and medical structures lack caregivers. This paper introduces an innovative and generic vitals monitoring software. This real-time and embedded software, called BCG-VARS, uses ballistocardiography to perform contactless measurements of vital parameters on various types of equipment (e.g. medical bed, car seat, wheelchair). The software consists in two algorithms and was designed to extract three different vital parameters: actigraphy, Breath-to-Breath Interval (BBI), and heart Inter-Beats Interval (IBI). The algorithms were designed to analyze ballistocardiography signals but can be reused for other pseudo-periodic signals. BCG-VARS shows state-of-the-art high performances, even when sensors are deeply integrated in a medical bed. Absolute mean errors of 1.59 beats per minute for IBIs and 1.03 cycles per minute for BBIs were obtained in the deep integration configuration. In addition, all the subjects presences and movements (lie down, roll over or leave) were detected by the algorithms. Thanks to parameter tuning and interval prediction, the system is accurate, reliable in multiple applications and adapts to every new individual. [Display omitted] •BCG measurement of vitals monitoring in real-time, without learning requirement.•Reusable algorithms for the majority of pseudo-sinusoidal signals.•Use of interval prediction on heart pattern, reducing large measurement errors.•Parameter tuning can be done, to fit a use case and improve interval prediction.