Learn AnaTele from scratch¶
This is an illustrated course. It assumes no background: you don't need to have flown a drone, opened a log or seen the maths before. By the end you'll be able to read an AnaTele report and explain every number in it.
Every chart is drawn from real data: flights from AnaTele's built-in simulator, analysed by the
same code you run. The script that draws them is
docs/figures/make_guide_figures.py.
Some pages also have small interactive widgets that run in your browser.
Prefer paper?
The same material is available as a printable PDF: AnaTele Illustrated Guide (PDF).
Reading map¶
| # | Chapter | You will learn |
|---|---|---|
| 1 | Drones 101 | Motors, roll/pitch/yaw, why a quadcopter can hover |
| 2 | Logs and MAVLink | What a flight log is and what is inside it |
| 3 | Time and sources | Clocks, reboots and why AnaTele uses epochs |
| 4 | Flight phases | How AnaTele finds takeoff, flight and the crash window |
| 5 | Attitude and tilt error | "Did the drone do what it was told?" in one number |
| 6 | Motors and mixing | How four motors make every movement, and saturation |
| 7 | Vibration, EKF and compass | Sensor health signals and what bad looks like |
| 8 | Rules and their four states | Thresholds, persistence and "unknown is not healthy" |
| 9 | Hypotheses | Turning evidence into a ranked list of explanations |
| 10 | Missions and geometry | Waypoints, distances on a sphere, "where did it go?" |
| 11 | The simulator | The physics model that made every chart here |
| 12 | Sharing with .ant files | One tamper-evident file for a whole investigation |
| Glossary | Every term in one place | |
| Interactive widgets | All the widgets on one page |

How to use this course¶
- In a hurry? Read chapters 1, 4, 5 and 9. That's enough to read a crash report.
- Investigating a real crash? Do chapters 4–9 next to the Diagnostics reference.
- Writing your own rules? Read chapters 8 and 9, then Writing a rule.
The one idea to keep in mind
AnaTele reports observations and hypotheses. It never gives a crash verdict or a probability of cause. If data is missing, the answer is unknown, never fine.