Designed Excitation Biased Towards High Frequencies

For a few realistic measurements I have conducted with a significant source of noise and a wide frequency range (about 2 decades) I have noticed that the designed excitation trends towards high frequency features during later iterations and misses or fails to resolve the lower frequency features. This is for a more complicated model to fit (8th order). I'll provide an animation to evaluate the evolution of the designed excitation over a few iterations as well as the data stored from this run.

I cannot upload my data here however it is stored on the controls computer at: /users/michael/sysID/data/srm_measurement_result_081226.npz

Start by obtaining the animation and the referenced dataset at /users/michael/sysID/data/srm_measurement_result_081226.npz, then reproduce the later excitation iterations for the 8th-order model. Compare the frequency coverage across iterations and identify why lower-frequency features are missed; done means the cause is documented and the excitation resolves the reported low-frequency behavior.

Written by the indexing model from the issue text.