Acoustic Nonlinearity Parameter Estimation for Exoskeleton Control

The acoustic nonlinearity parameter of skeletal muscles changes with their contractions and has the potential to be utilized for exoskeleton control. In our previous work, a system comprising a wearable ultrasound probe was developed to estimate the acoustic nonlinearity parameter
in vivo
by measuring amplitudes of second-harmonic ultrasound. The estimating accuracy was found to be significantly affected by the corruption between echoes due to the limited axial resolution of ultrasound. As both frequency components of the single echo and the extent of corruption are unknown, a method based on a bi-directional deconvolution is proposed in this paper to decompose corrupted echoes, and integrated into the measuring system. Eight subjects were recruited for the
in vivo
measurement of the biceps brachii acoustic nonlinearity parameter, thickness, and elbow torque. Linear regression between the elbow torque and (1) the estimated acoustic nonlinearity parameter, (2) the thickness demonstrated the correlations, with average coefficients of determinations (
$R^{2}$
) of 0.861 and 0.633 respectively. This work paves a way for estimating the acoustic nonlinearity parameter of skeletal muscles
in vivo
.