sim_flight_kalman.py

Run a synthetic rocket flight through a Python mirror of the AURORA Kalman filter and plot the result.

What it does

The script is a 1:1 Python port of the firmware filter pipeline. The two mirrored components are:

  • The two-state (altitude, velocity) constant-acceleration Kalman filter from aurora/lib/filter/kalman.c (Filter, filter_init, filter_predict, filter_update, filter_detect_apogee).

  • The body-frame gravity tracker from aurora/lib/sensor/attitude.c (Attitude, attitude_init, attitude_calibrate_sample, attitude_calibrate_finish, attitude_update, attitude_is_calibrated).

Function names, struct layout, return semantics (0 = OK, 1 = gated/apogee, negative errno on error) and arithmetic order all track the C, so any tuning change made in Kconfig or in the C sources can be reproduced in Python by updating the mirrored constants at the top of the file.

The script generates its own synthetic data:

  • A pressure signal from the ISA barometric formula.

  • A body-frame IMU stream (accel + gyro) with a slow pitch-over after apogee, so the gravity tracker actually has to do something.

  • MS5607-flavoured Gaussian noise on top of the pressure stream.

These streams are fed through the filter mirrors at 50 Hz and the results are handed to plot_flight_data.py, which produces the seven-panel plot shown below.

Note

Plotting is delegated to plot_flight_data — sim_flight_kalman.py no longer plots anything itself. To plot recorded telemetry from a flight log, use plot_flight_data.py --flight DIR directly.

Usage

python3 tools/sim_flight_kalman.py [--theme {light,dark,both}]
                                   [--show] [--title TITLE]

  • --theme {light,dark,both} — render plots matching the Furo Sphinx docs theme (default both).

  • --show — open the plots in an interactive matplotlib window in addition to writing them to disk.

  • --title TITLE — override the plot title.

Filter and attitude tuning is not exposed on the CLI on purpose: the constants at the top of the script (CONFIG_FILTER_*, CONFIG_IMU_UP_AXIS_*) are the same Kconfig defaults the firmware uses, so a default run is directly comparable to a default flight build. To sweep tuning, edit those constants or use sweep_apogee.py.

The output file is flight_simulation.png (or _light / _dark suffixed when --theme both is used). A typical run looks like this:

Ground ref pressure: 101326 Pa (1013.3 hPa)
True apogee:         500.0 m (t = 12.71 s)
Filter apogee:       499.7 m (t = 12.78 s)
Attitude g_mag:      9.794 m/s²
Gated baro updates:  0
[light] Plot saved to flight_simulation.png
flight_simulation_light.png flight_simulation_dark.png

Reuse from other scripts

The filter and attitude mirrors are exposed as plain module-level functions so they can be driven from any other Python tool that wants to replay or sweep them:

  • Filter, filter_init, filter_predict, filter_update, filter_detect_apogee, filter_votes

  • Attitude, attitude_init, attitude_calibrate_sample, attitude_calibrate_finish, attitude_update, attitude_is_calibrated

  • simulate_flight(...) — generate the synthetic trajectory used by the default run.

Helpers for loading recorded telemetry (parse_influx, parse_state_audit, segment_flights, etc.) now live in plot_flight_data.py.

Requirements

  • Python 3.10+

  • numpy, matplotlib