Hi, I’m a big fan of Arducopter and have it installed on several drones. I’ve been thinking of professional applications so attended Drone-X in London a few months ago. One of the companies there said they are switching from Arducopter to PX4 because the latter has “a more stable release cycle”. This saddened me since I know how powerful Arducopter is and how much I enjoy using it.
I’m writing this not wanting to cause any sort of ruckus. How could I “sell” the safety of Arducopter to potential industry partners? Is there such a thing as a “stable” version of Arducopter?
Again, I’d like to emphasise how appreciative I am of the efforts by all of you to make Arducopter so useful.
I very much doubt that the company based their decision on evidence, but would be interested in discussing it directly with them. With around 100 partner companies, and many other adopters that aren’t partners, I think that the commercial and enterprise use of ArduPilot is well proven. ArduPilot runs a 6 month release cycle, although beta testing/release hardening means that this isn’t strictly enforced (ie the team aims for 6 months but prioritises stability of the release over timelines).
I wish I remember the conversation better since it only just came to mind. It struck me as odd at the time as like you say, there are so many companies already using Arducopter and I’ve seen how each release is thoroughly tested.
Thanks for the reassurance.
Sorry for the second post. I just checked my notes and I think they said that Arducopter will never be safety certified so they had to switch to PX4. I believe these were the words used. As I say, safety-certified by whom is the question. And I do not know how well informed their decision was. Different laws for different countries. I’m talking about the UK here so I wonder if most of Arducopter support is outside the UK, which is probably a small market anyway.
if we speak about strict Avionic certification: neither Px4 or ArduPilot can reach it.
About Design verification and drone certification, I don’t think either flight stack will make any difference. And it is already proven that ArduPilot is used on safety certified drones (and even more) that operate all around the world in numerous context : patrolling, SAR, urban deliveries, etc.
Re certification, I’m guessing that the FAA were satisfied: Drone Delivery System Achieves FAA Type Certification | Unmanned Systems Technology
GitHub - matternet/ardupilot: Matternet modifications of diydrones/ardupilot
UK requirements will of course be different, but there’s drone deliveries happening everyday there too, powered by ArduPilot. Unfortunately there’s a lot of heresay/confusion in the drone game, some fueled by ignorance, some by competition or other motives. As a project we try not to get distracted by it, and focus on doing the best we can for our community. I’m glad you’re enjoying ArduCopter
Regarding support, one of the funded support engineers for the Partners Program is UK based, as are a number of other devs and partners. Support in the UK is as good or better than anywhere (apart from Australia, but we’re just lucky down here).
This is music to my ears, especially as I invested so much time in Arducopter.
I’m starting to wonder if it was some vested interest that got them to change autopilot software. The company that changed from Arducopter was dependent on another company for business/funding so it makes you wonder if they were dictating requirements.
This company might’ve also been sold on the “enterprise” aspect of PX4 and the entire auterion suite. QGC and auterions automatic flight uploads/reviews make things prettier and less of a pain for a lot of end users.