Treadmill velocity calculation

[Paulo Lucareli]

Dear all,

We are evaluating treadmill running and gait kinematics with a 3d optical motion capture system (8 Vicon 1 Mp cameras).

To calculate spatiotemporal parameters, it is necessary to determine the rolling belt speed. For this purpose, two circular markers 2d were glued onto the rolling belt. Even with the limitations of 2d marker reconstruction, it was possible to obtain data for analysis of the rolling belt speed. Constant variations in velocity were observed due to the loads applied by the subject. It is also possible that there are belt deformations and other sources that produce oscillations in speed.

We would like to know how researchers who analyze patients or athletes on treadmills determine the instantaneous speed displacement of the running belt or calculate the spatiotemporal parameters of movement.

Best

[Edmund Cramp]

Re: Treadmill velocity calculation

I have just been working with a researcher who needed to calculate the subject's stride length on a treadmill. After looking at some of her data in a C3D file I found it was reasonably easy just looking at the heel marker co-ordinates to determine the initial contact treadmill speed. Once that's known I think that all you need is the frame count to the next heel strike to calculate the stride length.

[Paulo Lucareli]

Re: Treadmill velocity calculation

I have just been working with a researcher who needed to calculate the subject's stride length on a treadmill. After looking at some of her data in a C3D file I found it was reasonably easy just looking at the heel marker co-ordinates to determine the initial contact treadmill speed. Once that's known I think that all you need is the frame count to the next heel strike to calculate the stride length.

If we use the forefoot marker to measure the speed of the treadmill we will have some error in the calculation, won't we? Especially if the initial contact occurs with hindfoot. That's why we think of using the treadmill belt but when the runner hits the belt the speed almost triples by impact
Velocity.jpg

Best regards

[Wagner De Godoy]

Dear Biomec-L subscriber,

Some treadmill running exams consistently showed variation in belt speed. In the figure from the previous post (Dr. Paulo Lucareli) the speed graph shows a peak in the double swing phase - the belt speed is decreased under dynamic loads. Perhaps only a few classes of equipment have variations in speed. Additional studies would be needed to verify this hypothesis.

A draft calculation, with the generation of a local coordinate system on the treadmill, was developed with a set of three reflective 2d markers attached to the belt treadmill. Two videos with a brief description are available on YouTube (motion analysis on the treadmill). Routines in the BodyBuilder programming language and c3d files (pre- and post-processing) are attached to this post: trial_04_hiae_3.zip (1).


The calculation model was divided into 3 routines (2) due to some processing failures - perhaps the programming can be redone and improved, whether in BodyBuilder or Matlab and Python.

(1) c3d files and speed graphics can be viewed in the Mokka software.

(2) software used to capture and process routines: BodyBuilder 3.6.4 and Vicon Nexus 1.5.1
processing sequence:
a - BodyBuilder - macro: 5e_field offset.mod / save c3d
b - macro:esteira - 5f_fieldoffset.mod / close c3d file
c - In Nexus 1.5 apply Woltring filter (the markers P1, P2 and P3 were excluded by previous processing of the "Fill Gaps" command)
d -BodyBuilder - macro: esteira-mkr_gait - 5g_field offset.mod
note: *.mp files can be read and edited on MS Windows Notepad

Note: BodyBuilder function
BODYBUILDER FOR BIOMECHANICS: Special Functions (page: 116)
<<
[ ] is a special "field offset" post-fix function used to find the final value of an expression in any field, relative to the current one.
segment P[-1] segment P in previous field

The "sample offset" function has many uses, including the creation of special filters.
>>

I apologize for the simplified explanation and the translation flaws. All criticisms about the mathematical model or programming logic will be useful.
Thank you for your attention.
Best Regards,

Wagner de Godoy
Brazil
(lasciate ogni speranza, voi ch'entrate)

[Wagner De Godoy]

Dear Biomec-L subscriber,

New paper about "treadmill velocity calculation":

Willwacher, S., Oberländer, K.D., Mai, P. et al. A new method for measuring treadmill belt velocity fluctuations: effects of treadmill type, body mass and locomotion speed.
Sci Rep 11, 2244 (2021).

https://doi.org/10.1038/s41598-021-81951-9

"Abstract
Treadmills are essential to the study of human and animal locomotion as well as for applied diagnostics in both sports and medicine. The quantification of relevant biomechanical and physiological variables requires a precise regulation of treadmill belt velocity (TBV). Here, we present a novel method for time-efficient tracking of TBV using standard 3D motion capture technology. Further, we analyzed TBV fluctuations of four different treadmills as seven participants walked and ran at target speeds ranging from 1.0 to 4.5 m/s. Using the novel method, we show that TBV regulation differs between treadmill types, and that certain features of TBV regulation are affected by the subjects’ body mass and their locomotion speed."


Best Regards,
Wagner de Godoy
Brazil