Continued work from klog31907.

Abstract:

I diagonalized modal actuator for GAS filters and it seems to work well.
Then, I tested modal damping control with decouled modal actuators and confirmed that the low frequency mode can be damped without inducing large noise at high frequeny.
Now, I implemented the modal damping of the GAS filters for the lowest and the second lowest mode into guardian, so it will automatically engaged at LOCK_ACQUISITION state.

Detail:

I decoupled modal actuator for GAS filters of ETMY.
Figure 1-5 show the TFs of M1-M5 excitation to M1-M5 sensors (blue: before decoupling, red: after decoupling).
All DoFs seem to decoupled well.

Then, I tested the modal damping of M1 and M2 because M1 and M2 signals have a peak at resonant frequency while the others don't have.
Figure 6 shows the GAS filter signals with modal damping.
Blue shows the spectrum at LOCK_ACQUISITION state without modal damping, green line shows the one with M1 modal damping without actuator decoupling, and the red shows the one with modal damping after actuator decoupling.
After the decoupling, noise induced at high frequency seems drastically reduced.

Figure 7 shows the spectra of ETMY OpLev pitch signals (the color represents the same configuration as fig6).
Same as GAS filter signals, high frequency noise is reduced from the yesterday condition, so actuator decoupling seems to work well.

Since induced high frequency noise is reduced and almost close to the spectra without modal damping, I implemented M1 and M2 modal damping control into the guardian.
So, these two modal damping controls are automatically turned on when ETMY goes to LOCK_ACQUISITION state.

Note:

In fig4, there is a peak at 0.8 Hz in M4 to M4 transfer function but I'm not so sure what it is; the resonant frequencies of 5 modes are about 0.2, 0.6, 1, 1.7, and 2 Hz, which are differennt from 0.8 Hz.