Question about inverse dynamics

[Tim Zulj]

Hey,

I'm beginner when it comes to OpenSim, so i have one question about inverse dynamic.

So my question is, what is the point of using force plates for inverse dynamics? I mean you can get all the information for inverse dynamic from kinematic model, at least in theory, isn't that right?

[Thomas Uchida]

what is the point of using force plates for inverse dynamics? I mean you can get all the information for inverse dynamic from kinematic model, at least in theory, isn't that right?

Yes, in theory, but kinematic data is likely to be noisier (particularly problematic if it's being differentiated) and less reliable (e.g., if your kinematic model assumes a rigid HAT segment). A couple resources with additional discussion:
- Kuo (1998) "A Least-Squares Estimation Approach to Improving the Precision of Inverse Dynamics Computations", https://doi.org/10.1115/1.2834295
- Hicks et al. (2015) "Is my model good enough? Best practices for verification and validation of musculoskeletal models and simulations of movement", https://nmbl.stanford.edu/publications/ ... ks2015.pdf

[Kaiwen Yang]

Hello Thomas,

I have additional question about ID. In OpenSim, why not just run RRA based on IK data to get joint torques?

Kaiwen

[Thomas Uchida]

why not just run RRA

The purpose of RRA is to reduce residuals. There are scenarios in which the residuals are not expected to be low (e.g., if the residuals stand in for unmeasured external forces). There are also applications in which iterating with RRA would be impractical, such as in this paper: Pizzolato et al. (2016) "Real-time inverse kinematics and inverse dynamics for lower limb applications using OpenSim", https://doi.org/10.1080/10255842.2016.1240789.

[Kaiwen Yang]

Thanks