# Эксперимент: State-Range Observer Gain Scheduler (#557)

{
  "worked": false,
  "confidence": 10,
  "verdict": "Built a reproducible two-state noisy observer with kernel projection, gain candidates, Lyapunov-margin selection, parameter sweeps, and fixed-gain comparisons. The core math check found C U_K=0 exactly, making U_K^T G C U_K identically zero for every gain; the observed projected stability boundary was gamma≈0 within the 0.01 grid resolution and did not move with rho. The scheduler selected rho=1 for stable cases and only fell back to rho=4 for unstable cases without fixing instability; on the stable noisy toy, scheduled RMSE matched fixed low gain (0.007893) and was better than fixed high gain (0.010423), but this is not evidence of the claimed mechanism.",
  "metrics": {
    "baseline": "Fixed low rho=1: tail RMSE 0.007893, max error 2.2061; fixed high rho=4: tail RMSE 0.010423, max error 2.1457.",
    "idea": "Scheduled gain: tail RMSE 0.007893, mean rho 1.0, max error 2.2061; projected-gain invariance error 0.0 and observed boundary error <=0.01."
  },
  "how_to_run": "/home/maxwelhelp/main/bin/python3 observer_scheduler_experiment.py",
  "files": [
    "observer_scheduler_experiment.py",
    "results.json"
  ],
  "limitations": "This intentionally tests the formulas as written in a linear two-state observer rather than a neural ODE or Lorenz-63 system. It does not solve an SDP for P, estimate Jacobians with autodiff, test hysteresis, or evaluate a corrected observer formulation where injection acts outside ker(C)."
}