I optimized MICH and PRCL feed forward.
Followings are what I did for optimization (basically same manner as in klog24334).

1. Measure the TF from MICH out to DARM without FF.
2. Measure the TF from ITMX TM_SUMOUT_L_OUT to DARM.
3. Calculate the ratio between MICH2DARM and ITMX2DARM (optimized MICH FF).
4. Measure the TF from MICH out to PRCL with MICH FF.
5. Measure the TF from PRCL out to PRCL.
6. Calculate the ratio between MICH2PRCL to PRCLout2PRCL (optimized MICH2PRCL FF).
7. Measure the TF from PRCL2DARM.
8. Set filters of ITMX inverse filters based on the measurement in 2.
9. Set PRCL2DARM filter based on the measurement in 7.

Measured data are stored at /users/Commissioning/data/{MICH/PRCL}/2023/1226/.

Note:

Somehow, PRCL2DARM has low coherence (fig1) though we can clearly see the peaks during the measurement.
It might be necessary for further optimization after stabilizing IFO.