Contactless Monitoring of Muscle Vibrations During Exercise with a Chaos-Inspired Radar

TL;DR

This work introduces GigaFlex, a contactless mmWave radar system that estimates muscle force and detects fatigue during exercise by exploiting chaos theory. By integrating a two-stage entropic localization, determinism-based segmentation, and chaos-informed learning (including demographic embeddings and few-shot calibration), the approach reconstructs muscle output from broadband, nonlinear vibrations. In a 23-participant study, it achieves MVIC RMSE of $5.9\%$ and fatigue-detection AUCs of $0.83$–$0.86$, performing comparably to a contact-based IMU baseline and outperforming wearables in fatigue detection. The results demonstrate the feasibility of noninvasive, real-time physiological sensing of non-periodic biosignals with potential for injury prevention and personalized training, while outlining steps toward broader deployment and multi-muscle sensing.

Abstract

In this paper, our goal is to enable quantitative feedback on muscle fatigue during exercise to optimize exercise effectiveness while minimizing injury risk. We seek to capture fatigue by monitoring surface vibrations that muscle exertion induces. Muscle vibrations are unique as they arise from the asynchronous firing of motor units, producing surface micro-displacements that are broadband, nonlinear, and seemingly stochastic. Accurately sensing these noise-like signals requires new algorithmic strategies that can uncover their underlying structure. We present GigaFlex the first contactless system that measures muscle vibrations using mmWave radar to infer muscle force and detect fatigue. GigaFlex draws on algorithmic foundations from Chaos theory to model the deterministic patterns of muscle vibrations and extend them to the radar domain. Specifically, we design a radar processing architecture that systematically infuses principles from Chaos theory and nonlinear dynamics throughout the sensing pipeline, spanning localization, segmentation, and learning, to estimate muscle forces during static and dynamic weight-bearing exercises. Across a 23-participant study, GigaFlex estimates maximum voluntary isometric contraction (MVIC) root mean square error (RMSE) of 5.9\%, and detects one to three Repetitions in Reserve (RIR), a key quantitative muscle fatigue metric, with an AUC of 0.83 to 0.86, performing comparably to a contact-based IMU baseline. Our system can enable timely feedback that can help prevent fatigue-induced injury, and opens new opportunities for physiological sensing of complex, non-periodic biosignals.

Figures (12)

• Figure 1: GigaFlex continuously monitors muscle fatigue and force to help users maximize workout effectiveness while minimizing the risk of injuries with long recovery periods.
• Figure 2: (a) Neuromuscular origins of surface muscle vibration. (b) Experimental setup measuring muscle force output for the isometric (static) exercise of elbow flexion. (c) Similarity between muscle vibrations at the bicep measured using a contact-based accelerometer and at a distance of 40 cm using a mmWave radar.
• Figure 3: Muscle vibrations are broadband and look similar to body motion in the (a) time and (b) frequency domain.
• Figure 4: Recurrence plots for different scenarios: static isometric exercise, dynamic isotonic exercise, and noise. These plots show how similar a muscle signal is to itself over time. Dark cells indicate when there is a repeated pattern.
• Figure 5: (a) Muscle localization via entropy. Classical radar localization using CFAR produces false positives from body movement (knee movement) as well as motion in the environment (dumbbell motion), and can only be used as a coarse-grained filter. The Chaos dynamics of muscle vibrations can be characterized using an entropic "sieve", $\sigma_{ENTR}$, where values below a certain threshold can be discarded, and the maximum value can be regarded as the muscle vibration signal. (b) Segmentation of lifting phase via determinism. Relying on radar displacement to determine the start and end of a lift is unreliable because the signal is often contaminated by small tremors and subtle body motions. In contrast, determinism (DET) metric captures the temporal regularity of muscle vibrations, enabling more precise segmentation of the lifting phase.