litchi mission almost drove my P3 into a cliff

Joined
Aug 23, 2014
Messages
76
Reaction score
21
Website
www.stevenmichelsen.com
I used Litchi to create a mission over some mountains - gimbal pitch disable and heading on manual, so that I can point the camera as the mission proceeds. I used the "above ground" feature so that every waypoint would be the same height AGL. I used a LOT of waypoints, close together, as I didn't want to get caught crashing into some pointy bit of rock. I checked the mission by exporting a KML to Google Earth to confirm clearances.

Apparently all that was not enough - see the video. I bailed out of the mission at literally THE LAST SECOND.

I now have to wonder - is it more accurate to create this sort of mission "the old fashioned way", using Google Earth as the basis for the waypoints, and then import into Litchi? I was under the impression that the data using either method comes from the same place, and that the results would be the same. Now I wonder...

Please let me know your impressions about all this. Did I not use enough waypoints?

I edited the elevation AGL on a lot of the waypoints after this experience and flew the mission without incident, though it surely seemed like I was closer to the ground than I expected to be in a lot of places. Hmmmm…...

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
I edited the elevation AGL on a lot of the waypoints after this experience and flew the mission without incident, though it surely seemed like I was closer to the ground than I expected to be in a lot of places
I have not looked at the video as of yet, however, looking only at the sample graphic above, did you have signal loss during the mission at any point? Your vertical altitude lines are not linear. This could be part of your editing but I am asking the question anyway. Also, the barometer in the Phantom is not that accurate, nor is Google Earths representation of such. Allway's add at least an additional 20 feet above your planned altitude, just for the error in the barometer reporting. If you can pull the aircraft .dat file for this flight we can most likely tell you where the issue lies. Also, when posting issues with Litchi, please include a link to the mission profile from the hub. Lastly, if you decide to look at the additional data and are not familiar with the process you will need to retrieve the aircrafts's .dat by following the instructions on the link below. Then share a link back here to that. You will need to upload the retrieved file to a share site such as Drobox, Google Drive etc... to be able to share that here. They are quite large files. Link is below if interested.

Retrieve Dat File
 
I think Google Earth elevations are not super accurate, and may be more generalized across an area. In the video, you can see the hillside suddenly jut upward in the form of a rock outcrop. I'm guessing the hillside may continue it's slopped ascent beyond that outcrop as your waypoints reflect, but either Google Earth did not pick up the sudden increase in terrain at that point, or a waypoint just didn't happen to fall there and Litchi just assumed a constant slope between.
 
  • Like
Reactions: athearnsd45
60 feet is only 18.3 m AGL I moved the path of one of my missions that was 18m AGL from a valley to a ridge line and my P3 didn't make it over the highest point (soft landing into a canola crop luckily )
The way GE ground elevation are measured results in an average over a 30m radius apparently so sudden changes in ground elevation will not be well represented.
I checked the GE elevation of a TRIG station near us and this point was 11.5m higher than GEs figure .
As FD said add 20 feet at least then review your footage carefully to work out where GE is not telling you the truth.
 
60 feet is only 18.3 m AGL I moved the path of one of my missions that was 18m AGL from a valley to a ridge line and my P3 didn't make it over the highest point (soft landing into a canola crop luckily )
The way GE ground elevation are measured results in an average over a 30m radius apparently so sudden changes in ground elevation will not be well represented.
I checked the GE elevation of a TRIG station near us and this point was 11.5m higher than GEs figure .
As FD said add 20 feet at least then review your footage carefully to work out where GE is not telling you the truth.

Clearly I put too much trust in the elevations provided. I was not aware of that 30m radius average. That's a big deal when dealing with outcrops and such! I THOUGHT I added enough extra waypoints to avoid missing such features, but given the averaging that GE apparently does, in that case no matter how tightly I packed the WPs, they would be working with averages and not actual elevations. UGH.

FWIW here's my mission. The outcrop is at WP 13. Mission Hub - Litchi

I tweaked the mission and re-flew it with no problems after that almost fateful trip. I will post that video, and also the .dat file that fly dawg requested when I can.
 
Last edited:
I still need to uncover the .dat file, but for now, here's the edited flight - the second trip, all AGL numbers increased by at least 80 feet on location. Considering that the AC seems to fly really low from time to time in this video, it makes me think that a LOT of the numbers on the original mission were way off. Enjoy:

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
 
You were lucky you were able to stop it the first time while still in range .If it had cleared the first obstacle there were plenty more lined up. Very nice scenery and an ambitiousmission.
 
I have found that after rising over 100ft. or so, further along a mission the reported altitudes are off, sometimes over 50ft. Is that Aridzona?
 
Hello @Steve_in_NJ.
@The Springs sent me a link to this thread.

OMG, Trees are more forgiving then those rocks. ;)
Glad you didn't have to retrieve it!

Your shared mission you posted doesn't come up, did you change it?
Could anybody else view it?

I'm looking through this thread kind of in a hurry.
If you didn't loose signal than your mobile device has a good log.
crApples are a pain to retrieve the files from, I use an iPad, iPhone but not a Mac so I'm always frustrated. (You have to use iTunes)

Downloading the .dat files is not fun either, but if you haven't done it before with a working bird, it is even more frustrating.

I would like to see one of your missions or both.

I really wish @VegasFlyer would show up again, its been since June!
My most daring mission!

Rod
 
Last edited:
@Steve_in_NJ

Your posted mission doesn't come up, Would really like to see it.

Rod
 
I used Litchi to create a mission over some mountains - gimbal pitch disable and heading on manual, so that I can point the camera as the mission proceeds. I used the "above ground" feature so that every waypoint would be the same height AGL. I used a LOT of waypoints, close together, as I didn't want to get caught crashing into some pointy bit of rock. I checked the mission by exporting a KML to Google Earth to confirm clearances.

Apparently all that was not enough - see the video. I bailed out of the mission at literally THE LAST SECOND.

I now have to wonder - is it more accurate to create this sort of mission "the old fashioned way", using Google Earth as the basis for the waypoints, and then import into Litchi? I was under the impression that the data using either method comes from the same place, and that the results would be the same. Now I wonder...

Please let me know your impressions about all this. Did I not use enough waypoints?

I edited the elevation AGL on a lot of the waypoints after this experience and flew the mission without incident, though it surely seemed like I was closer to the ground than I expected to be in a lot of places. Hmmmm…...

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.
Well, looks like I can say I just found a guy who's much braver than I. Flying a P3 in a complex waypoint mission in Litchi would have given me 'Buns of Steel'. I have a P4P v2 and I am very careful when I set up my waypoint missions. I check them top to bottom very carefully before flying them. But I would still have crashed my drone twice if I had not had Obstacle Avoidance. Once was when my drone came across a 300' radio tower that did not show up on the map I was using. It had been installed very recently.

Another time was when I was headed into some pesky power lines about a mile away. OA saw them and stopped me but a P3 wouldn't have. And seeing those lines on my iPad screen in time to manually stop would have been nearly impossible.

Overall, Obstacle Avoidance has saved me thousands of dollars in the 2 years I've had my DJI P4P. Once was during an attack by a red-tailed hawk. Funny to watch that predator overshoot his target because the drone would stop in midair when it "saw" the hawk trying to take a swipe at it.

The most obvious time disaster was averted was when I set up a Litchi Orbit of a 60' monument. The P4P would lift off and head straight for the monument at 17mph and stop within a foot or so of crashing into it. I could not figure out why and tried again. Same result. Finally spotted the problem. The bird was trying to enter the Orbit circle from the opposite side of it's path. Changed setting to 'Nearest' and all was good....but a P3 would have been totally destroyed. OK, rant's over and my hat's off to intrepid Steve.
 
Last edited:

Recent Posts

Members online

Forum statistics

Threads
143,066
Messages
1,467,352
Members
104,933
Latest member
mactechnic