Robot-supported assessment of balance in standing and walking

Table 4 Review of metrics used to quantify balance performance during standing and walking, usable for robotic assessment

Definition of metric	Suitability for Walking or Standing	References
AREA of COP or COM motion (e.g., 95% confidence circle or ellipse area)	Standing	[17, 86,87,88]
PATH LENGTH of COP or COM motion	Standing	[86, 87]
STATISTICS of COP or COM motion (e.g., centroid, x^th percentile, median, dispersion)	Standing	[17]
DISPLACEMENT of COP or COM (e.g., RMS, maximum excursion, range, ratio of AP/ML)	Standing and Walking	Based on COP [17, 19, 89,90,91] Based on COM [92]
RELATIVE MOTION of COM relative to COP; COM to BOS; COP to BOS (e.g euclidian distance, time-to-contact)	Standing and Walking	[93,94,95,96,97]
SYMMETRY of COP or COM motion (e.g., left-right ratio)	Standing and Walking	[98,99,100,101,102]
LONG-TERM CORRELATION in COP or COM motions (e.g., Hurst exponent, Largest Lyapunov Exponent)	Standing and Walking	[103,104,105,106,107,108]
VELOCITY of COP or COM (e.g., average, range)	Standing and Walking	[17, 19, 86, 88, 90, 109]
TOTAL BODY ORIENTATION (e.g., range of angle or angular velocity)	Standing and Walking	[89]
JOINT KINEMATICS - angle or angular velocity of joints (e.g., RMS, maximum excursion, range, time series)	Standing and Walking	[110]
JOINT KINETICS - joint torque, torque rate of change (e.g., RMS, maximum excursion, range, time series)	Standing and Walking	[76]
TEMPORAL-SPATIAL gait parameters (i.e., step length, stride length, step width, swing duration, temporal stance-swing ratio)	Walking	[76]

The table lists metrics that quantify balance, as used in posturography and gait analysis. Most of the presented metrics have been validated by showing that they are significantly different among groups or conditions with different balance behaviour such as elderly versus young adults, or eyes-opened versus eyes-closed conditions. Readers are referred to the indicated studies for approaches or methods for calculation, and for detailed information about which groups or conditions the methods have been shown to be indicative.
These metrics are all applicable to assess all three types of balance control: steady state, anticipatory, and reactive. It can be expected that descriptive statistics e.g., average, minimum, or maximum of these metrics are highly influenced by the assessment procedure, especially during reactive balance, when reacting to device-specific perturbations.

ISSN: 1743-0003