Who/How/Why -- do you trust non-DJI automation apps? (Litchi, DroneLink, DroneDeploy, etc...)
- Thread starter Firejay27
- Start date
dronepilotbob
Well-Known Member
My one and only crash since 2014 was using mavic air with litchi app. I haven’t used it since. I have tried Dronelink and DroneDeploy without any issues after the fact.
To be fair I’m 99% confident that the air had a compass error and it was pilot error since I should have waited for more satellites to lock in (regardless of HOME being locked) and I should have checked my true north. I’m blaming this one on myself and not litchis fault. Regardless, superstition has me staying away from it.
To be fair I’m 99% confident that the air had a compass error and it was pilot error since I should have waited for more satellites to lock in (regardless of HOME being locked) and I should have checked my true north. I’m blaming this one on myself and not litchis fault. Regardless, superstition has me staying away from it.
I use Litchi pretty much exclusively and never have a problem except by my own doing. Litchi is solid. There are some cosmetic changes in the device display I would like to see though.
I am using Litchi, dronelink and several Mapping apps. I do not see a risk in using those apps. As long as you keep an eye on the flight and are ready with the pause or Sports mode button to stop auto flight at any time if needed.
Worth mentioning is also the new EU rules that probably will allow for an extended VLOS as long as you are using preprogrammed flight. The thought behind that is that the pilot can then scan the sky more easily for traffic. This is from my memory after reading some standard operations proposals some weeks ago. Needs to be verified!
Worth mentioning is also the new EU rules that probably will allow for an extended VLOS as long as you are using preprogrammed flight. The thought behind that is that the pilot can then scan the sky more easily for traffic. This is from my memory after reading some standard operations proposals some weeks ago. Needs to be verified!
I was "flying sideways"
Like 45° to the right towards the reveal so it got "sideswiped"
Actual footage:
first angle adjustment at 10 sec is flying straight to the start, then the second angle adjustment at 14 sec is the angle to the right of straight ahead.
A good feature they can add is “site visit” before the mission. Basically an always “head forward” submission to make sure the path is clear.
That is what you can do in Virtual Litchi Mission.
I would have cleared it if Google topo were correct.
1st 1/4 is the planned mission.
The rest is what I had to climb - the mission pointed straight forward
The saddle at the top is where I hit yet the flight path says I would clear it
i didn’t mean “virtually “. A real site run before the real mission which may have saved your drone
. Because in the site run your drone will face forward.I too have flown many Litchi missions without problems. The error with Google maps is scary though. In generating animations with Google Earth I have encountered altitude errors when crossing the division line of 2 of the strips Google uses to map the terrain. In these instances, the errors were around 30 feet, judging from the impact on the animation. There is always going to be some error in the ground altitude as reported by Google as well as any other mapping program. The question is how much. Perhaps someone with more geospatial knowledge can tell us what a safe margin would be.
The same type of question come up with the accuracy of GPS or GLONAS positioning for waypoints. The GPS specified accuracy is within +- 16 feet 95 % of the time. Laying out a route tighter than that means a greater than 5% change the drone will hit an object on either side of that waypoint if it is high enough.
The same type of question come up with the accuracy of GPS or GLONAS positioning for waypoints. The GPS specified accuracy is within +- 16 feet 95 % of the time. Laying out a route tighter than that means a greater than 5% change the drone will hit an object on either side of that waypoint if it is high enough.
So far as what I have seen, the Spark has been right on the money.
The Mavic Air, not so much...
Planned route here was to follow the road it started on, go around the town square and continue on as it did as seen:
Some reason it started to veer off at the first 3 waypoints and continued to veer off.
Because it corrected itself as it did, there was a weird gimbal action and the visual jerk when complete.
Here is a side by side mission, same flight plan for both - Spark on the left:
See the Spark execute perfectly around Cross Mountain but the Mavic Air swings way out around the back side.
Not off with the GPS, there are reports of Litchi calculations not 100% with Mavic Air
I had 2 Sparks and 2 Mavic Air - Sparks were consistent as well as the fault with the Mavic Air being consistent
Understanding how the application works in detail, and careful experience. The software always does what is programmed to do. The DJI api also is fixed, allowing only certain action sequences. These are the same actions that 3rd parties call.
The next things to consider are the error margins for the input data. GPS has one set of probabilistic errors which will be different each time you fly due to satellite drone geometry. Google earth has errors based on the individual surface strips used in making up the composite surface
Litchi uses both of these sources to calculate its instructions to the drone. Errors in the data input will cause slightly incorrect instructions out of Litchi.
Litchi also has 2 quite different modes for dealing with waypoints. The first is what navigators refer to as shoelace mode, where each leg of the flight represents a straight line between the 2 waypoints. The 2nd mode is interpreted, where it calculates a flight path which is optimized around a path between multiple waypoints to provide a smoother flight. This rarely gets to any single waypoint but flies a graceful curving path through the sequential waypoint field. Either choice also incorporates the data input errors.
I have flown many Litchi flights with P3 4K, MA, and P4adv with no calculated errors. I have seen several errors where cross track errors occur probably due to GPS imprecision, never greater than 10-15 feet.
The next things to consider are the error margins for the input data. GPS has one set of probabilistic errors which will be different each time you fly due to satellite drone geometry. Google earth has errors based on the individual surface strips used in making up the composite surface
Litchi uses both of these sources to calculate its instructions to the drone. Errors in the data input will cause slightly incorrect instructions out of Litchi.
Litchi also has 2 quite different modes for dealing with waypoints. The first is what navigators refer to as shoelace mode, where each leg of the flight represents a straight line between the 2 waypoints. The 2nd mode is interpreted, where it calculates a flight path which is optimized around a path between multiple waypoints to provide a smoother flight. This rarely gets to any single waypoint but flies a graceful curving path through the sequential waypoint field. Either choice also incorporates the data input errors.
I have flown many Litchi flights with P3 4K, MA, and P4adv with no calculated errors. I have seen several errors where cross track errors occur probably due to GPS imprecision, never greater than 10-15 feet.
I agree 100%. "How do I trust other apps?" I don't, not at all. However, I DO trust STATE FARM. I bought an insurance policy for my Mavic Pro. It's $60 dollars, US, per year. Total coverage, $1,700
So if Litchi looses my drone, I'll end up with a new one! Kind of makes me hope it happens. however, in 2 years, flying at least weekly, still no Litchie Glitchie.
dronelink
Well-Known Member
- Joined
- Dec 3, 2018
- Messages
- 153
- Reactions
- 181
- Age
- 41
️?Similar threads
DJI Drone Deals
New Threads
-
-
-
-
-
Remembering the Zombie Apocalypse Drone Footage- Started by The Droning Company
- Replies: 2
Forum statistics
