Welcome to SkydioPilots.com
Sign up and join the discussion!
Sign up

Skydio 2 crash (yeah!)

stefanonwheels

Active member
Joined
Dec 17, 2019
Messages
40
Reaction score
24
I had my Skydio 2 follow me on my mountain bike ride in the Arizona desert and my Skydio 2 crashed. When trying to stay in front of me, it flew towards the sun backwards and probably didn't see the ocotillo it then flew right into.
Yes, the sun was getting low and yes, the branches of an ocotillo are not very thick, so the conditions weren't optimal. However, if I can't fly my Skydio 2 in Arizona in winter, when the sun is starting to get lower at around 3 PM and when ocotillos don't have leaves, that'd be a bummer and limit the range of usability for that drone quite a bit as there's tons of other situations with similarly challenging light & obstacle conditions.
Luckily, it landed on rather soft ground and was mostly unharmed, except for some very light scratches on the lenses of 3 of the 6 obstacle avoidance cameras. I flew the drone right afterwards and it was fine. I'm worried whether the scratches on the lenses will cause future obstacle avoidance issues, though, so I will contact Skydio support about the crash.

I think this shows what the Skydio 2's limits are, though, so the great (and maybe blue-eyed) feeling to be able to fly this drone anywhere with full confidence that it won't crash is a thing of the past now and, well, back to reality. I guess it's time to try out Skydio's US-based support now and to see how customer-friendly they are when it comes to warranty claims.

Video on YouTube:
 
Last edited:
Sounds as if S2’s OA suffers from the same into the sun troubles as the Intel RealSense. I’m not even sure if a LiDAR system could overcome those conditions. The sun is just too bright!
 
Sounds as if S2’s OA suffers from the same into the sun troubles as the Intel RealSense. I’m not even sure if a LiDAR system could overcome those conditions. The sun is just too bright!

Yeah this video is more telling. Having GPS means the sun didn't matter for XY positioning. Altitude control is okay. I think obstacle avoidence failed maybe due to the sun. Light and dark will always be a problem for optical (maybe IR lidar..) and flying at speed it's difficult to see twigs, etc..

I'd say it's better than a human pilot who would suffer from the same conditons. From experience I set my altitude above the trees/bushes. Hopefully SD software will mature to negotiate these conditions or fly above them.
 
Yeah this video is more telling. Having GPS means the sun didn't matter for XY positioning. Altitude control is okay. I think obstacle avoidence failed maybe due to the sun. Light and dark will always be a problem for optical (maybe IR lidar..) and flying at speed it's difficult to see twigs, etc..

I'd say it's better than a human pilot who would suffer from the same conditons. From experience I set my altitude above the trees/bushes. Hopefully SD software will mature to negotiate these conditions or fly above them.

See the Skydio can't do the test at <10 mph you want me to do at 30 mph with the M2P. I'm tempted to try though, since I have a new drone in the way and Care Refresh!!! Hahaha!
 
Hey Blue Bird I won't waste any more time with this guy, it's clear his agenda.

Lon I own a P1, P3P, P4, Mavic Pro Mavic Pro 2.....
 
Hey Blue Bird I won't waste any more time with this guy, it's clear his agenda.

Lon I own a P1, P3P, P4, Mavic Pro Mavic Pro 2.....

What "agenda"? I own drones from 5 different companies myself. DJI can go out of business tomorrow, and I wouldn't care in the least. I just want to get some cool shots of bridges and cranes.
 
Sounds as if S2’s OA suffers from the same into the sun troubles as the Intel RealSense. I’m not even sure if a LiDAR system could overcome those conditions. The sun is just too bright!
Well, nobody expects OA to work flawlessly at all times. What I do expect, though, is for an AI-powered drone to know its own OA's limits and to stop flying if it isn't 100% sure that the path is clear.
If Skydio 2 had stopped, hovered and made the beacon beep to get my attention, I would've been 100% happy with that.
 
  • Like
Reactions: DoomMeister
Well, nobody expects OA to work flawlessly at all times. What I do expect, though, is for an AI-powered drone to know its own OA's limits and to stop flying if it isn't 100% sure that the path is clear.
If Skydio 2 had stopped, hovered and made the beacon beep to get my attention, I would've been 100% happy with that.

It thought the path was clear that is why it crashed, same things happen to humans all the time!
 
  • Like
Reactions: Ridefreak
Hey Blue Bird I won't waste any more time with this guy, it's clear his agenda.

Lon I own a P1, P3P, P4, Mavic Pro Mavic Pro 2.....
No worries, Lon will always promote what he knows best, however he still does not have the S2 in hand to do a full and accurate comparison. Sorry Lon.
 
  • Like
Reactions: stefanonwheels
No worries, Lon will always promote what he knows best, however he still does not have the S2 in hand to do a full and accurate comparison. Sorry Lon.

Geez, I'm like who is "Blue"? Hahaha! That's true, supposedly it's on the way. They sent me some of the accessories though. So just to be fair, it's my understanding the he doesn't have a Skydio 2 either so if I can't make a "fair and accurate comparison" neither can he! Sorry Robbie... At least I have something...20191225_134819.jpg
 
Skydio support finally got back to me with the expected disappointing response:
"Regarding your crash, Thank you for uploading the logs and sending the video. We are glad that your Skydio 2 survived that tumble. I'm sure you have gone back seen how many close calls you had with thin branches. While Skydio 2 is a giant leap forward in autonomous tracking there are still limitations. While that type of collision would not have been covered under the warranty, there are steps you can take to avoid future crashes. For example, if you had set the Skydio 2 to a higher altitude and maybe further away it would have avoided most if not all the thin branches completely. While I don't think you should have to anticipate every trail you're flying in, It's always beneficial to take some time to consider the situations you are placing your Skydio 2 in. It's always better to err on the side of safety when it comes to co-piloting with the Skydio 2. You do have some incredible footage on your channel and we really hope that Skydio 2 can continue to help you capture amazing footage."

Skydio support also mentioned that they don't even have the means to repair Skydio 2 drones at this point, so even if I was willing to pay them, they wouldn't be able to fix my drone at this point. So right now, if your camera lens is scratched, well, please buy another $999 drone.

Here's my response to Skydio support:
"Hi John,

got it. Thanks for confirming the disappointing but expected response. I won't even comment on it. Could you please just convey the following to your product development team?

In all the situations that we've seen Skydio 2 crash so far, would the R1 have crashed? Are you really making a step forward in usable (!!!) drone autonomy?
I sincerely hope that you don't seriously believe "innovation" is to sell a drone whose significant limitations require you to enforce a customer-unfriendly warranty statement that rules out about 50% of the areas that people would like their autonomous drone to be able to fly in.
Failure to acknowledge deficiencies in a product is the beginning of the end of that product, I hope you realize that. People are less and less capable, expect more and more and the world consists of thin branches and power lines, so please make your drone able to cope with those situations, otherwise is pretty much worthless to me and many other people who use Skydio 2 for what it's made for.

Stefan"
 
  • Like
Reactions: parkgt
Forewarned is forearmed I guess. Thanks to Stefan for sharing. I guess a lot of the other guys whose drones wouldn't fly after their crashes will be out of luck. I fly several drones. I crash all of them. My Mavic pro has taken four tumbles. The highest was from 25 feet onto a mowed lawn under a tree. I have seen it land successfully on half a propeller. I had to fix the gimbal with my fingers once (and it still works well). The camera has actually improved after the crashes. So far it has proven very tough. My f11 needed an arm glued after a much lower fall. I have a cheap plastic Heron toy drone. The Heron I believe is the answer that Skydio needs. It is made out of a flexible light weight plastic. It is light weight. It has built in prop guards. One day I had it in the back of my truck and some cans of pop got on top of it. It was deformed for a few days but eventually popped back into shape. Because of this, I have crashed it from 20 feet onto concrete with absolutely no ill effects. I believe it could fall from any height. The downside is that there must be 0 wind to fly it. Or, it is like a kite. What if they used a hard plastic inner coreUntitled.png to give it weight and built most of it out of flex plastic? That's the answer. I would put money on it.
 

New Posts

Members online

No members online now.

Forum statistics

Threads
1,600
Messages
12,368
Members
2,363
Latest member
ViktorSinatorov0fup