How long it took this recovery window to get back inside the hover band.
Position Error
0.00 m
Attitude Error
0.0 deg
Wind Force
0.0 N
Settling Hint
Recovering
Controller
LQR
Route Mode
Off
Largest position miss during the current recovery or route segment.
Average gap between the true vehicle and the Kalman state estimate.
Largest control burst used to recover, useful for LQR vs PID comparison.
Last LQR Run
Run the guided demo to capture a clean LQR recovery sample.
Last PID Run
Switch to PID or use the guided demo to generate a baseline.
Comparison Verdict
Capture both an LQR run and a PID run to generate a direct recovery comparison.
Motor Command
Estimated vs True Error
Control Effort
Altitude / Route Progress
Start the controls walkthrough
This guided run stages a clean comparison sequence: LQR recovery, disturbance rejection, PID baseline behavior, then route tracking.
Shortcuts
R Reset
G Gust
M Mode
T Target
O Route
Space Pause
D Demo
E Export
The true dynamics are integrated nonlinearly with quaternions, while the controller and estimator operate on a hover linearization. That separation keeps the physics expressive while letting you feel how optimal control and state estimation work in practice.
- Rotor mixer with saturation and realized torque reconstruction
- Optional PID baseline for side-by-side controller feel
- Motor spool lag and route playback for less toy-like behavior
- Live Q/R tuning for controller and filter behavior
- Noisy sensor fusion for position, attitude, and angular-rate estimates
What to notice
LQR should usually settle faster with a smoother effort story because it is optimizing a full state-error cost. PID is a strong baseline, but it reacts more locally and tends to overshoot or work harder when the same gust hits.