Open AccessUpper body estimation of muscle forces, muscle states, and joint motion using an extended Kalman filter
Author(s) -
Mohammadi Hanieh,
Khademi Gholamreza,
Simon Dan,
Bogert Antonie J.,
Richter Hanz
Publication year - 2020
Publication title -
iet control theory and applications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.059
H-Index - 108
eISSN - 1751-8652
pISSN - 1751-8644
DOI - 10.1049/iet-cta.2020.0321
Subject(s) - joint (building) , kalman filter , estimator , standard deviation , control theory (sociology) , torque , computer science , mathematics , physics , engineering , structural engineering , artificial intelligence , statistics , control (management) , thermodynamics
This study synthesises modelling techniques and dynamic state estimation techniques for the simultaneous estimation of the muscle states, muscle forces, and joint motion states of a dynamic human arm model. The estimator considers both muscle dynamics and motion dynamics. The arm model has two joints and six muscles and contains dynamics both of the muscles and of the motion. We develop an optimally tuned extended Kalman filter using noisy measurements of joint angles with standard deviation 2.87 ° , of joint velocities with standard deviation 6.9 ° /s, and of muscle activations with standard deviation 10% of their peak values, and then simultaneously estimate joint angles, joint velocities, muscle forces, joint moments, and muscle states. The standard deviations of estimation errors (SDEE) are no more than 0.07° for joint angles, 1 ° /s for joint velocities, 0.6 mm for muscle–tendon lengths, and 0.1 Nm for joint torques. The results are compared with a previously developed static optimisation method, and verify the effectiveness of the proposed estimator in providing lower SDEE for both muscle and motion dynamics of the human arm model compared to the static optimisation method.