Fig 1.
Structure of the motor actuated lower-limb orthosis (MALO) exoskeleton corset, frames, and shoes with single degree-of-freedom rotation joints.
Hip and knee joints are equipped with direct current motors.
Fig 2.
(a) The subject is tripped by an obstacle in the middle of the walking lane. A safety harness was connected to the gondola to support the subject in the cases of falls. Gait timing is controlled using a guide pole and metronome. (b) The reaction force of the obstacle was used to detect tripping and switch the assist pattern to reaction algorithms. The assist torque was applied based on the gait cycle estimated from the timing of gait events.
Fig 3.
Typical reaction motion against tripping.
Stick picture of the recovery motion in the sagittal plane. The left leg was tripped in this figure. The subject stepped the tripped leg forward. Then, the recovery leg was stepped.
Table 1.
Median and interquartile ranges of tested parameters.
Table 2.
Correlation coefficients of studied parameters.
Fig 4.
Factor loading of each principal component.
The three major principal components which were associated with the respective proportions of variance of 22.3, 18.9, and 16.2%.
Fig 5.
Boxplot of reaction parameters and principal component parameters which were affected by the reaction algorithm (p < .05) were extracted.
Box represents the first and third quartiles. Outliers, which are separated from the quartile more than 1.5 times the interquartile range, are marked with crosses.
Fig 6.
Step motion of recovery strategies.
The recovery motion could be classified based on the length of recovery steps. (a) The first step after tripping was large, whereas the second step was small. (b) The first step was small and the second step was large.