Teach-in 2001 : CoM & CoP during Rhythmic Weight-shifting
by Chris Kirtley, The Catholic University of America, Washington DC

One fundamental component of the control of locomotion is the ability to shift weight from one leg to the other. This is done by using the Center of Pressure (CoP) to drive the total body Center of Mass (CoM). In this experiment, a subject was asked to shift body weight alternately from one foot to another in time with a metronome. Both feet were on seperate force platforms, so that the individual and net CoPs could be measured. The CoM was also calculated (at 60 fps) from a full body marker set (Eames et al, 1999), and is shown by the long vertical green line. The CoP is the small inverted 'T' on the floor. You can see that they move sinusoidally pretty much in time with each other, but the amplitude of the CoP is a little larger. Be patient loading the animation as it's about 1.8Mb.

Winter (1995) suggested that the the CoM is related to the CoP by

d(CoM) = k (CoP-CoM)
dt2

If the (CoP - CoM) difference is calculated (the CoMCoP:x graph, low-pass filtered and inverted for clarity), you can see that it does indeed resemble the CoM acceleration signal (CoMAcc). If we plot CoMAcc/(CoP - CoM), we get:

It looks like the value of k (the gain of the postural control system) is hovering around -10, but there are big spikes rather regularly spaced through the data. Click on the chart to download the Excel file (500kB) and have a closer look.

Questions
• What is causing these strange spikes?
• Why are they regularly spaced?
• How could you get rid of them?
• Does k really represent the gain of the postural control system?
• Would you expect k always to be around -10?
• in antero-posterior motion?
• in different subjects?
• in subjects of different heights?
• in subjects of different body mass?
• What does k represent in biomechanical terms?