Fig. 3

Predictive targeting algorithm control flow diagram. The targeting algorithm receives the desired percent swing input from the experimenter and the Fy,Fz, and Mx signals from the instrumented treadmill. Once the experimenter triggers a perturbation (a), the system waits until the next toe-off event (b), then passes to the time delay block where the time delay, trelease, is received from the Targeting Algorithm. The system then releases the obstacle following the time delay (c) which results in a targeted perturbation during the subject’s swing phase. The unit step plots indicate the obstacle release signal at each point in the flow diagram: (a) indicates the immediate switch to high at the time of the trigger, (b) indicates the delay of the switch due to the time between the trigger and the subsequent toe-off, and (c) represents the added algorithmically-computed time delay before the obstacle is released. Note that trelease is calculated in Eqs. (1) – (10)