NFT @ 02102026: Towards warm-up readiness from multivariate wearable physiology in HIFT

#researchatnecstlab - #NECSTFridayTalk On Friday, 2 October, 2026, we had a new talk for the series #NECSTFridayTalk. During this talk, we had, as speaker, Susanna Bardini, PhD student at Dipartimento di Elettronica, Informazione e Bioingegneria. In the following, you can find the details about the talk: 📌 Title: Towards warm-up readiness from multivariate wearable physiology in HIFT 📌 Abstract: Physiological readiness at the end of a warm-up is typically inferred from the completion of a predefined protocol rather than from direct physiological evidence, despite substantial inter-individual variability. Since physiological adaptation during warm-up develops progressively without clearly defined boundaries, readiness is better conceptualized as a latent state. This talk presents a probabilistic framework based on a Hidden Markov Model (HMM) to identify readiness from physiological data collected with a medical-grade wearable device during general warm-up sessions preceding CrossFit training. Feature analysis of heart rate, electrodermal activity, and skin temperature shows that instantaneous measurements alone are insufficient to characterize warm-up progression, as similar physiological values may occur at different stages depending on the preceding physiological history. By introducing history-dependent features, the model reliably discriminates between warm-up phases within a compact three-state HMM. A structured initialization of the emission distributions further stabilizes state decoding across sessions. Readiness is ultimately defined probabilistically as the posterior probability of the Warmed state exceeding a predefined threshold for a minimum number of consecutive windows. These results support the use of wearable physiological monitoring for an objective, probabilistic assessment of readiness for exercise. #NECSTLab #Computerscience