Low/zero constraint reaction forces during IAA

[Lydia Brough]

I'm running an induced acceleration analysis on a walking trial using a rolling constraint. I've adapted the Hamner et al. (2010) setup file, which has worked for me in the past. (Using the updated rolling constraint setup from the 4.1 documentation gives an ''Array index out of bounds'' error.)

My results (below from 4.1) have the total vertical COM acceleration = gravity + noise, and the constraint reaction forces are close to zero. It seems like the constraint isn't really kicking in, even if I set it to being enforced as the default. I've tried the analysis in 3.3 and 4.1 and with a few different models (gait2354, Hamner, Rajagopal). Using the same experimental data and setup file in 3.3, the constraint reactions and COM acceleration were really really high (like 40,000N). The setup works fine on gait2354 sample data in 3.3, but not my data and I haven't been able to identify any major differences. I'm using data from the Matlab c3d converter, which has worked great through RRA and CMC. The states, GRFs and COP data look fine visually, but I wonder if there's something that could be giving IAA trouble. GRFs were filtered in Matlab but unfiltered data straight from the c3d file gives similar results.

IAA_COM_Y.PNG (17.9 KiB) Viewed 254 times

IAA_constraint_rxns.PNG (21.66 KiB) Viewed 254 times

I've attached the setup file and I can send other files if needed. I greatly appreciate any troubleshooting advice!