A sensor-driven Hill-type muscle modeling framework integrating sEMG and pFMG for biceps brachii force estimation.

Abstract

Hill-type muscle models offer physiologically interpretablemuscle force estimation,
but their use in wearable and personalized applications is limited by sensing constraints and
validation challenges. This study presents a sensor-driven Hill-type muscle modeling
framework for estimating biceps muscle force under controlled isometric conditions using
wearable signals. A Hill-type muscle model integrating surface electromyography
(sEMG) and pressure-based force myography (pFMG) was proposed to estimate biceps
brachii force under isometric conditions. To evaluate the mechanical consistency of the
model, a purely mechanical elbow model, based on joint geometry, moment arms, external
loads, and static equilibrium, was employed. The proposed Hill-type muscle model estimates
both passive and total muscle forces across multiple elbow angles, with results evaluated for
mechanical consistency against the benchmark mechanical elbow model. 
Under passive conditions, pFMG measurements alone enabled estimation of
length-dependent passive muscle force. For active conditions, the combined use of
sEMG-derived activation and pFMG-derived deformation allowed consistent estimation of
total muscle force trends. Strong agreement was observed between the sensor-driven
estimates by the Hill-type model and mechanically derived reference values by the elbow
model, with high coefficients of determination and low estimation errors. By
explicitly separating neural activation and geometric deformation within a Hill-type
structure, the proposed approach provides a physiologically meaningful and experimentally
feasible solution for wearable muscle force estimation, and may offer a potential foundation
for future investigation of real-time and subject-specific neuromuscular modeling
applications.