I’m researching reliability and crash investigation in UAV systems and wanted to understand real experiences from operators and developers working with Pixhawk-based drones.
A few questions I’m curious about:
1. When a drone crashes, how often are the onboard logs incomplete or corrupted?
2. Have you ever lost a drone because the crash location couldn’t be determined accurately?
3. How reliable are Pixhawk / ArduPilot logs when diagnosing the root cause of failures?
4. What tools or workflows do you typically use to investigate crashes?
5. What is usually the hardest part of analyzing a drone failure?
6. How much time does a typical crash investigation take?
7. If a drone goes down in a remote area, how do you usually recover it?
8. Do operators commonly add external GPS trackers, beacons, or recovery devices?
I’m trying to understand the real operational pain points around UAV failures and incident analysis.
Would really appreciate hearing about your experiences.
In about a dozen crashes, I basically never had a corrupted log. The last part (a few seconds) went missing though. The main question is if your data storage is physically intact.
Using GPS and having telemetry helps enormously for this, so note down the last known position shown in QGC. Alternatively, FPV pilots often use a beeper to find their drones.
So far, I never lost one because i could not find it.
The main limitation is the user who is looking at the logs. The logs are phenomenal, but not trivial to understand.
Analyze the logs with px4review or plotjugger, and compare with other available records.
Missing Logs. Or vague descriptions of the incident.
An easy one: 2 minutes. A hard one: Two days. If its user error, it is usually faster, if there is an actual bug in the code, it may be more complex.
I can only speak for myself, but if i fly further away, I would always have a plan on how to recover it.
