Non-Asymptotical Postural Stabilization Strategy during Human Quiet Stance

Postural sway, i.e., center of pressure and center of mass (CoM) and electromyograms (EMGs) from ankle muscles were measured during human quiet stance. It was confirmed that EMG activity of ankle extensors correlated with the CoM and CoM velocity, as reported earlier. However, the EMG activity was phasic and burst-like, compared to slow and smooth changes in CoM. It was demonstrated that the phasic EMG activity tended to appear only at specific periods mostly when the acceleration of CoM was close to zero and to terminate when CoM velocity changed its sign from positive to negative. We claimed that a simulated human stance model with a conventional PID feedback controller could not account for such phasic EMG activity. The phasic EMG activity was then associated with the active ankle torque, which is defined as the residual component of the stiffness torque in the total ankle torque. To this end, the active torque was estimated by assuming that the ankle stiffness was not large enough to make the upright posture asymptotically stable. The active torque and the phasic EMG activity were sharply correlated, implying that the phasic EMG activity might generate the active torque, and it might be produced based on a strategy other than the conventional PID feedback control