Adventures abound.

Programmed a Litchi Mission to go through the Frenchman Mountain Trail.

Sent it up because after the 2nd waypoint was going to be a reveal of Las Vegas.

Apparently there is a Google Maps altitude error and the Spark could not clear the crest.

It was 1250ft above home at that point so the Spark altitude limit was not a factor.

So guess who had to go up and get it?

Well, do not trust Google Maps.
Map resolution is vertical resolution of 30 meters and a horizontal resolution of 90 meters.
Considering that slope uses both resolutions, that’s a huge margin for error.

Hope little birdy is unharmed

Darn.
You got it back in 1 piece?

How did you retrieve it?

--- For unknown places, I always mark my waypoints flying there and hitting "waypoint at aircraft"

Sorry to watch that
Did you get it back? Any broken part?

Yep, sometimes  it is really good to fly/create some waypoints with "set waypoint at aircraft"  on the customizable buttons.
I "gave up", more or less, to fly very close to objects b/c of  just the GPS differences on the same mission, whe flown forward and backwards, sometimes!
Or if flown with two drones at the same time, but still theu not flying 100% the same course!
(I still work on that matter, with drone videos drone ;-)

Did you pre-check your mission with "virtual litchi mission"?
it is really helpfull at hillsides

Thanks for all the comments.
Got it back by climbing that mountain.
Used a hiking app with the exact coordinates shown to find it.
Fortunately it was close (20ft) to the trail and not on some ridge line.

Haven't tested it yet but IMU and Compass are in the green and it boots OK without any gimbal errors.
Props look sound.

I ran the mission through VLM several times and it should have cleared.
I took an altitude reading at the crest whee it crashed and compared it to home point and I got 1255ft
Litchi altitude in the ( ) next to the WP altitude said 1441ft so it had plenty of clearance.

Crazy thing about it, it should have been grossly over compensated. That altitude was grabbed by a ridge line on the right.
I curved the mission so the Spark would not pause at WP2 making a smooth transition.
The curve was directly over the more shallow crest.

I am impressed with the mission planning you guys do. I'm at the stage where I take my Spark with me and, if an opportunity presents itself, I launch and take some shots. Then, if my Spark comes back and lands successfully, I count myself lucky and move on hoping for another opportunity.

I take it, from the way you located your Spark, Litchi doesn't have "find my drone" yet.

Warriewoodlad Posted at 12-31 14:23
I am impressed with the mission planning you guys do. I'm at the stage where I take my Spark with me and, if an opportunity presents itself, I launch and take some shots. Then, if my Spark comes back and lands successfully, I count myself lucky and move on hoping for another opportunity.

I take it, from the way you located your Spark, Litchi doesn't have "find my drone" yet.

Litchi has "Find my aircraft", last known position shown in map view. But, no offline map. Sadly.

pmshop Posted at 12-31 13:37
Thanks for all the comments.
Got it back by climbing that mountain.
Used a hiking app with the exact coordinates shown to find it.

Great news PM!
Now, if you was giving it another try, what would you do differently? About the mission planning I mean..

Haggi Posted at 12-31 14:42
Great news PM!
Now, if you was giving it another try, what would you do differently? About the mission planning I mean..

IF, which I am not, but if I was, I would add another 100ft to WP 2 & 5.
I would plug in the crash coordinates as the WP 2 & 5 with no curve and deal with the pause there because the reveal is not until after the crest.
Run it through VLM several times, angled and straight views as before and try it.

Oh, and get there hours before the flight and post a sign "Have you seen me" and post the flight path.
For the hikers to keep an eye out
I ain't doing that s again!!!

Warriewoodlad Posted at 12-31 14:23
I am impressed with the mission planning you guys do. I'm at the stage where I take my Spark with me and, if an opportunity presents itself, I launch and take some shots. Then, if my Spark comes back and lands successfully, I count myself lucky and move on hoping for another opportunity.

I take it, from the way you located your Spark, Litchi doesn't have "find my drone" yet.

Thanks,

I come from a long background of automated missions.
Ever since the time I flew my first mission 2/12/16 with my home made craft.
It amazed me what i could do with it.
One day I missed an opportunity shot when the compass would not register correctly on my home made.
By the time I fixed it, the opp was gone.
So I needed a reliable, on-and-go craft and chose the Spark.

Flew it manually a lot until the 3rd party automated software started supporting Spark.
It really is a blast hitting "GO" and seeing it do what you programmed it to do.

As far as the "find my drone" it was still pinging for 20 minutes after the crash.
I would drop the antenna while climbing, hear it call out "Warning. Disconnected", raise the antenna and get reconnected with 11 GPS sats showing.
So I got a screen shot.
Once I got to the top, I plugged in the coordinates to a hiking app and saw it was 90 ft away and zeroed right in on it.

Your matra about trying the missions is exactly like mine.
I started small:

Got bigger:

and bigger (too big)


Small steps gets you a long way...like 5100ft away and 1255ft up

pmshop Posted at 12-31 15:23
IF, which I am not, but if I was, I would add another 100ft to WP 2 & 5.
I would plug in the crash coordinates as the WP 2 & 5 with no curve and deal with the pause there because the reveal is not until after the crest.
Run it through VLM several times, angled and straight views as before and try it.

I think there is alot to learn from your accident.
If you dont mind, I want to discuss/share some Litchi mission stuff.
First of all, is it possible to get your mission file as kml file to make a quick check? And you spark track as kml too? If privacy is a prob, just PM me.

I’d love to send the files for you to check but it will but it will be well into the 2nd before I can.
Privacy is not an issue as it was a vacation spot

pmshop Posted at 2018-12-31 18:01
I’d love to send the files for you to check but it will but it will be well into the 2nd before I can.
Privacy is not an issue as it was a vacation spot

I'm curious if Litchi mission got some limitations.

Scenario: Roundtrip, roadside ascend to WP at peak, all curved turns..

Is Spark able to ascend in 45° up to peak? At any set cruising speed?
We know (via VLM) Litchi divide curved turns in 5 segments.
Lets pretend there was a 150m/450ft radius at the WP, then it will start turning toward next WP 60m/180ft before WP. Will it still ascend after turn is initiated?
Will it stay inside a safety corridor of..let say.. 10m/30ft when ascending? Even in turns?

These questions can maybe be answered by comparison of mission and actual track.

Good to see you got it back.

Oops.

I used to hang around with gyroplane and helicopter pilots. Their line was always, "If you don't want to land on it, don't fly over it."

It doesn't look like it hit hard, so it's probably ok.

Tentoes Posted at 1-1 12:10
Oops.

I used to hang around with gyroplane and helicopter pilots. Their line was always, "If you don't want to land on it, don't fly over it."

It seems to be OK.
I created this the next day in high winds:
Hangar 360 View - Henderson, NV

Haggi Posted at 1-1 02:28
I'm curious if Litchi mission got some limitations.

Scenario: Roundtrip, roadside ascend to WP at peak, all curved turns..

Here you go Haggi,

KML

Litchi CSV Mission

The second is the mission as flown.

pmshop Posted at 1-1 17:13
Here you go Haggi,

KML

Thank you for the files!
Unfortunatly I cannot make use of the csv file, it only contain a few poi's, no trail. Is it possible to go via Airdata (free account and sync of DJI logs) and export as kml?
You..sir..is a daredevil. Take it as a compliment!

Glad you recovered it. What speed was it flying at?

Haggi Posted at 1-2 00:08
Thank you for the files!
Unfortunatly I cannot make use of the csv file, it only contain a few poi's, no trail. Is it possible to go via Airdata (free account and sync of DJI logs) and export as kml?
You..sir..is a daredevil. Take it as a compliment!

Try this - KML from Drone LogBook

Drone LogBook KML

And AirData

AirData KML

BrianSC Posted at 1-2 00:44
Glad you recovered it. What speed was it flying at?

According to the logs, max speed was 26.37mph but programmed for 33.6mph

The video was 2x for 60fps upscale but returned to normal play speed about 10 frames before collision.

Wow the feature of making these waypoints is pretty impressive!  I would like to experiment with these sometime and your experience certainly helps understand how it works and what to watch out for!  

pmshop Posted at 1-2 07:54
Try this - KML from Drone LogBook

Drone LogBook KML

Hi PM!
According to Airdata file, little Sparky pretty much nailed the mission plan. To my pleasant surprise. You was just too low.

I do not see this as a glitch, as we cannot trust topo maps. Until the world is covered with high accuracy Lidar data, we all need to take some considerations into the mission planing.

Could the incident been avoided?
Well, Litchi give you the option to batch edit WP's.
Maybe if you did first run pointing drone towards next WP, 50m/150ft above ground? Check footage for ground levels/ obstacles.
Then do final run with drone pointing toward poi's and closer to ground.

Bottom line, I would probably do the same mistake in a simular situation.

Thanks for sharing your data!

Haggi Posted at 1-2 08:50
Hi PM!
According to Airdata file, little Sparky pretty much nailed the mission plan. To my pleasant surprise. You was just too low.

That is why I dropped the WP2 pin on the ridge line to the right of approach.
The Altitude for the curve should have been the same as 175ft above the ridge line.

ajmckay Posted at 1-2 08:26
Wow the feature of making these waypoints is pretty impressive!  I would like to experiment with these sometime and your experience certainly helps understand how it works and what to watch out for!

Any time you have questions just let me know.

that hurts

SAADHERO Posted at 1-2 11:24
that hurts

You have no idea with that hike up that mountain

Using Virtual Litchi Mission, this is what I was looking for of the reveal if the Spark had not crashed.
Also, what I had to climb to get the Spark back.

Unlucky mate,

I found out that Google maps had that crest at 3016 ft.
Actual was 3660.
So there ya go.

pmshop Posted at 1-13 09:08
I found out that Google maps had that crest at 3016 ft.
Actual was 3660.
So there ya go.

Maps in general cannot be trusted for elevation.
Infact, there are very few reliable sources for heights.
Thank you for the update!

I found myself yelling at the screen "pull up, pull up" but alas it didn't help. Sorry!!

Actually, after I hand controlled couldn't get my bird to get up a flank of a hill in the speed I could get to (fly into) the hill, I have played around with an litchi testing mission.
And got some results, which reminded me on this mission of pmshop. ;-)

Can it be that there was just not enough time, for the Spark, to climb to the needed height, because the speed was "almost" full speed (50kph)?

My playing around was a mission, setup in the field, several waypoints, different heights, speed over 20 kph
Not always was the Spark on the planed height, after reaching the waypoint.
And b/c there was another wp coming, it just continued climbing/descending from where it was

The "not so nice" thing is, litchi is going to the next waypoint, in almost the speed you are telling it to use.
But it is, sometimes, just not reaching the top altitude, in case the vertical speed would have to be be above the 3m/s - 9,8ft/s it can do, without assistant 2 modifications
There is no logical  check, afaik, so litchi is assuming during your planning, that the bird can do what you want it to do.

You "ordered" max speed, and that it could not do this is, imho, a sign, it tried to gain the height you wanted it to be at.
But failed, perhaps, in the end?

How about you look at it from this angle, again?

S-e-ven Posted at 9-7 05:05
Actually, after I hand controlled couldn't get my bird to get up a flank of a hill in the speed I could get to (fly into) the hill, I have played around with an litchi testing mission.
And got some results, which reminded me on this mission of pmshop. ;-)

Yes, considered...
But the logs had me at the proper altitude that I programmed in.
The fault was Google Maps had the crest at a lower attitude than actual.
I found they had the crest at 3016ft
The actual was 3660