On Laterally Perturbed Human Stance: Experiment, Model, and Control

Figure 10

Best (a) and worst (b) prediction on . Plotted are the ankle and hip angle in deg. over time in seconds. The red/dashed lines are the experimental trials, and the blue/solid lines are the corresponding simulation of controller . Both trials are qualitatively correct and the system is stabilized. The best prediction on the left shows a slight underestimation in the ankle and a slight overestimation in the hip at the end of trial, predicting the rest of the trajectory with high precision. The right (worst) prediction on the test set underestimates the experimental trajectory midtrial, while the beginning and end are predicted fine.

(a)

(b)