Virtual Litchi Mission

[namirda]

You know that feeling when you're about to ask a question that has a totally obvious answer that two days of playing with VLM has hidden from you? That's me!

a) Should the flight time in GE be the same(ish) as that shown at the top of Litchi Hub? Mine never are … GE is always a lot longer,
… and ...
b) Should the Cruise Speed be represented correctly in VLM (as it is in calculating that Litchi Hub flight time) because, with ALL WPs set to Cruise Speed I can still change the mission settings value, Mission Hub reflects that change, but almost no change is perceived in GE.

PS - this software is awesome! When those problems are sorted it will be a lot more awesomererer!

There are a few things to understand about cruising speed :

1) The csv file output by Litchi unfortunately does not contain the cruising speed which you set in the Litchi mission settings. If you examine the csv file you will see that a value of zero is output instead of the desired cruising speed. This means that in order to get a virtual flight which matches the real deal you need to tell VLM the cruising speed. You should set the cruising speed in VLM to be the same as the cruising speed set in Ltichi.

2) This issue goes away if you explicitly specify the speed for each leg of your mission. In this case the cruising speed is not used and VLM should fly the mission at the correct speed for each leg.

3) As you have noticed, the flight times calculated by VLM and shown in the Activity Log are often not the same as the flight times calculated by Litchi. I think if you do the arithmetic yourself with a calculator you will find that the VLM figures are correct.

4) A number of people have compared virtual missions with their real flights and they are generally pretty good all things considered. For example, in post 588 on this thread, Prismatic found a speed correction factor of 0.996!!

Hope this answers your questions

N

 

[namirda]

Thanks for the response. The resolution of the screen captures I sent in the previous message was 3840 x 2160. I just loaded VLM on another computer (Microsoft Surface Pro 2) with resolution set to 1920 x 1080 and the VLM "Settings" dialog box looks the same as the first computer (see attachment). Maybe it's my old eyes!

Thanks again for a fantastic tool. I'm having lots of fun with it. If you could add a control to raise the outside temperature and calm the winds it would be perfect!

Sorry Radio-Flyer - that must be really annoying. Both the weather and the fonts! You should come to NZ - the weather is perfect just now.

Does anybody else out there have a similar display problem? I wonder why it has not come up before.

N

 

[namirda]

Thanks for the response. The resolution of the screen captures I sent in the previous message was 3840 x 2160. I just loaded VLM on another computer (Microsoft Surface Pro 2) with resolution set to 1920 x 1080 and the VLM "Settings" dialog box looks the same as the first computer (see attachment). Maybe it's my old eyes!

Thanks again for a fantastic tool. I'm having lots of fun with it. If you could add a control to raise the outside temperature and calm the winds it would be perfect!

I have just been playing with this problem a bit and I see a problem similar to yours if I set the Font size too high.

Right click on the desktop. Choose Display Settings and set Font Size to 100% or perhaps 125%

Does that help?

N

 

[OzoneVibe]

There are a few things to understand about cruising speed :

1) The csv file output by Litchi unfortunately does not contain the cruising speed which you set in the Litchi mission settings. If you examine the csv file you will see that a value of zero is output instead of the desired cruising speed. This means that in order to get a virtual flight which matches the real deal you need to tell VLM the cruising speed. You should set the cruising speed in VLM to be the same as the cruising speed set in Ltichi.

2) This issue goes away if you explicitly specify the speed for each leg of your mission. In this case the cruising speed is not used and VLM should fly the mission at the correct speed for each leg.

3) As you have noticed, the flight times calculated by VLM and shown in the Activity Log are often not the same as the flight times calculated by Litchi. I think if you do the arithmetic yourself with a calculator you will find that the VLM figures are correct.

4) A number of people have compared virtual missions with their real flights and they are generally pretty good all things considered. For example, in post 588 on this thread, Prismatic found a speed correction factor of 0.996!!

Hope this answers your questions

N

1) - Thanks for that!
2) - 99 waypoint mission with changing speeds … but I hear what you say. Thanks!
3) - I'd go further - Litchi Hub isn't correct, I discovered, just last evening! Draw a straight - 2 waypoint mission of exactly 1 km (or 1 mile) and set speeds to 30 kmh (or 30 mph) - and it gives you a time of 3 minutes.
(I've used 999m in the screen shot to show that it's not 1.5km)

(And, don't wory, I'm not really going to fly down the local international airport runway @ 30m off the deck. )
I've checked other flight times in Litchi Hub, where speed is constant, and I'd agree yours are far better!

Thanks for your helpful response.

Edit: I heard from Litchi ….

 

[radio_flyer]

I have just been playing with this problem a bit and I see a problem similar to yours if I set the Font size too high.

Right click on the desktop. Choose Display Settings and set Font Size to 100% or perhaps 125%

Does that help?

N

Thanks namirda, changing the font scale in Windows display settings fixed the problem. I normally have scale set to 150% on the 4k monitor; when I change it to 100% the settings dialogue box looks like it should.

I had tried changing scaling before and it didn't make a difference. I wasn't exiting VLM though; I see I have to restart VLM for the change to take effect. I should have thought of that.

So you have the solutions for both my font and wx issues. Changing font is free but the wx fix is going to cost me!

Thanks namirda!

 

[parkgt]

What is the easiest way to get the GE Virtual Mission into a video editor?

 

[namirda]

What is the easiest way to get the GE Virtual Mission into a video editor?

Google Earth has a built-in movie maker - Go to GE/Tools/Movie-Maker.

Works perfectly

N

 

[parkgt]

Google Earth has a built-in movie maker - Go to GE/Tools/Movie-Maker.

Works perfectly

N

Thanks

 

[marklyn]

Thanks for the response. The resolution of the screen captures I sent in the previous message was 3840 x 2160. I just loaded VLM on another computer (Microsoft Surface Pro 2) with resolution set to 1920 x 1080 and the VLM "Settings" dialog box looks the same as the first computer (see attachment). Maybe it's my old eyes!

Thanks again for a fantastic tool. I'm having lots of fun with it. If you could add a control to raise the outside temperature and calm the winds it would be perfect!

I have a relative who has a higher end monitor setup (4k) and we are experiencing the same issue as described above. Text in window very small and 'bunched up" as shown. Also, when I move his mouse to grab a waypoint the cursor doesn't grab the point unless I'm a few centimeters off/away from the point. We tried another monitor (1920x1080) (dual monitors setup on the pc). No other apps/programs on either of these monitors have this issue, it's only with VLM. Google earth works fine too so I think the issue is some incompatibility between VLM and the way it displays with his video card (GeForce GTX970)

 

[OzoneVibe]

@namirda Hi - recently, I'm not always getting an Altitudes plot in the Profile Plot window at the bottom of VLM.
The Heading and Gimbal Angles plot is available every time.
Any ideas?
TIA!

 

[namirda]

@namirda Hi - recently, I'm not always getting an Altitudes plot in the Profile Plot window at the bottom of VLM.
The Heading and Gimbal Angles plot is available every time.
Any ideas?
TIA!

View attachment 55326

View attachment 55327

Hi,

Yes - I've seen this a couple of time as well. If you open the Developer Console (Tool Button top right of Profile Panel) you will probably find that it is a javascript error of the form "google.visualization.DataTable is not a constructor"

If you google that error message you will find that the same problem occurs randomly in many other applications as well as VLM. I don't think anybody has got to the bottom of it yet - I certainly haven't. For now it remains one of life's mysteries!

N

 

[Rockowe]

To view this content we will need your consent to set third party cookies.
For more detailed information, see our cookies page.

Accept third party cookies

Virtual Litchi Mission is a free application (kudos to Namirda!) that lets you fully automate a drone flight. I planned this flight over the summer of 2018, and was able to fly it on Sept. 21, 2018.

About the Location:
The video was taken in the American Southwest, a region that includes Arizona, Utah, Nevada, most of New Mexico, parts of Colorado, and a big hunk of southern California. I call this special site "The Spirit of the Painted Desert". It makes me happy to know that, somewhere out there, a place like this actually exists. I hope that makes you happy, too.

About the Technology:
Y'all get this, right? VLM puts a robotic precision pilot in your Mavic, and it expertly follows a highly detailed flight plan. You can layer on control of the heading and gimbal, to make everything but the exposure automatic. I guess you could leave that on Auto, too, and just have a Daquiri while your bird does its thing.

About the Flight:
The wind was brutal. I debated even flying. But the light was exquisite. Part of my kit was an identical "lite" tour that gave me heading & gimbal control. I could pre-fly the mission with positive look-ahead to ensure safety along the entire path. That test flight proved out, despite the wind, so I launched the true mission. And held my breath for 10 long minutes. The video speaks for itself, and Bertie came home.

Disclaimer:
For the record, I have no affiliation or relationship of any kind, but particularly not of a financial kind, with User Namirda (all I have for a name) or probably anyone he-or-she knows. Nor do I have any such relationship with VC Technology Ltd, the makers of the Litchi family of products.

And I honestly only bought Litchi (it's cheap) to give VLM a shot. I was, and remain, stunned by what this software combo makes not just possible, but simple!

TRULY AWSOME - Some Close Shaves there... but Magnificent -
Be Careful on path Curvature at waypoints - and Google smoothing -
Roamer lost his drone that way (click)

 

[Rockowe]

Download latest version here :
Virtual Litchi Mission V2.3.0

N

Hi Namirda,

Google Earth PITCH not interpolating Logarithmically - but rather Linearly (Hence inaccurate)

Iv just discovered what may be a major calculation Glitch - with Possibly VLC, G-Earth or Litchi for that Matter.

SO Heres the Scenario.

I figured My Drone (Mavic 2) was Quite Happy with a 12.5 RoundTrip.
… SO I marked the Points from Start (1) - to WP(2) about 5.5 km away.
Then another wp(2) at 2.4 - and finally home (wp3) 4.5km.

Something Like a THIN slice of Pizza [See The Simple 3x point mission] Click Link )

My Home is on a Mountain (about 80m in the air) - so all waypoints were at an altitude of 10m compared with takeoff -
Wp2 (5.5km away) - on the ground was about 100m below (Had a Restaurant - I knew well)
and
Wp3 (2.4km away) - on the ground was about 100m below (also Had a Swimming Pool - GYM)
before for turning for home … 4.5 km away

So I figured a Good Way to turn would be around Each Waypoint Feature - No Problem - I put POI 2 and 3 at these places (1m) above the ground)

OKAY now to test the Mission BEFORE filling up with Half way waypoints (minimum distances)

From Home to WP2 the View Looked Good... But something was off. It kept INCREASING the Pitching DOWN uniformally.

So at Half way (ie 2.8km from home and 2.8km from wp2) the Pitch was Almost at 45 Degrees.
It started at 0deg (home) and just LINEARLY started increasing till (POI 2 ie WP2)
Which Had a Waypoint closeby that was aiming almost -90deg (See Litchi Plan).

SO by the time I arrived at WP2 (POI2) the Gimbal Pitch was corrected again where the Linear increase had eventually found the Full down look - (ie -90deg)

BUT the interpolation IS NOT LINEAR

At home (5.5km away) - looking at WP2 the Gimbal would be say minus 2deg - and ONLY when the Drone was at 100m away - should the Pitch have been at -45deg (On Google Earth).

… but that's not how the Google earth presentation was. at HALFway it was at 45degrees. (Which was still 2.8km away from the WP2)

It means either Google EARTH and/or Litchi and/or DJI may be calculating this transition inaccurately.
I can see that google earth is not doing the Gimbal transition correctly - .

The FInal Waypoint should be in the Middle of the screen Up until the FInal Overfly.
The Calculation is reverse Logarithmic or ARCTAN
… not Linear

so if say X is the Far distance to WP2
and Y is the Altitude
and Aa = Angle between
… then TAN(Aa) = Y/X
then Aa = arctan((Y/X) Not a linear factor of Y

Im sorry to get so Glitchi Here … but it explains the inaccurate views see on VLM when running the mission google earth - (It kinda Gets it but doesn't)

Solution: Maybe to use this more accurately … it would be good when VLM writes the CSV into the Watch folder - it "Creates" included and correctly calculated sub waypoints - that represent the gradual Angular Views.

I tried to FIX this (Using the POI feature as apposed to Interpolate) but to no avail
… the interpolation of the GIMBAL is not mathematically correct.

The only way around this (bad fix) is to put Many Waypoints Approaching and leaving the Waypoint !!!

Maybe Iv got something fundamentally wrong here - … But Iv used VLM often and I LOVE IT !!!! - and was happy blaming Google for Inaccuracies.

See The Simple (essentially 3x point mission) Click Link
Tell me what you think...

 

[namirda]

Hi Namirda,

Google Earth PITCH not interpolating Logarithmically - but rather Linearly (Hence inaccurate)

...........

Tell me what you think...

Hi Rockowe,

I can't see your mission on the hub - perhaps because you have not marked it as public. I'll take a look at it when I can see it!

Thanks

N

 

[Rockowe]

Hi Rockowe,

I can't see your mission on the hub - perhaps because you have not marked it as public. I'll take a look at it when I can see it!

Thanks

N

Try it Now with Reference to my Long notes - (Made PUBLIC)

 

[namirda]

Try it Now with Reference to my Long notes - (Made PUBLIC)

Hi Rockowe,

OK - I've looked at your mission and I think I get your point.

In summary considering your WP1 and WP2 only:

1) Both your WP1 and WP2 use POI2.
2) WP1 has a gimbal pitch (input by the user) of -1 degrees and WP2 has a gimbal pitch (calculated by Litchi) of -79 degrees
3) The camera is properly focussed on POI2 at both WP1 and WP2

I think your problem is that the camera will NOT continue to point at POI2 as the drone flies from WP1 to WP2 - and you are correct about that!

For straight line legs between waypoints (i.e. ignoring the complication of curves near waypoints), your drone will simply perform linear interpolation of altitude, heading and gimbal pitch as it flies from one waypont to the next. That is the way DJI have programmed their firmware and that is how I mimicked it in VLM. Nothing fancy - simple linear interpolation.

In fact there is nothing else that DJI could have done because your drone does not know that WP1 and WP2 use the same POI - it does not know anything about POIs at all!!

POIs are a feature of the Litchi and VLM software only and are not transmitted to the drone. The drone only knows the altitude, heading and gimbal pitch at each waypoint and interpolates linearly between them.

In this case where two consecutive waypoints both use the same POI, I could, as you point out, do the interpolation differently so as to keep the camera pointing at POI2 for the duration of the leg - but since the whole point of VLM is to mimic the real flight there would be no point in doing that!

If you want to have the camera pointing continually at POI2 as you fly from WP1 to WP2, you will need to generate additional waypoints between them

I hope that helps.

Neil

 

[Rockowe]

Hi Rockowe,

OK - I've looked at your mission and I think I get your point.

In summary considering your WP1 and WP2 only:

1) Both your WP1 and WP2 use POI2.
2) WP1 has a gimbal pitch (input by the user) of -1 degrees and WP2 has a gimbal pitch (calculated by Litchi) of -79 degrees
3) The camera is properly focussed on POI2 at both WP1 and WP2

I think your problem is that the camera will NOT continue to point at POI2 as the drone flies from WP1 to WP2 - and you are correct about that!

For straight line legs between waypoints (i.e. ignoring the complication of curves near waypoints), your drone will simply perform linear interpolation of altitude, heading and gimbal pitch as it flies from one waypont to the next. That is the way DJI have programmed their firmware and that is how I mimicked it in VLM. Nothing fancy - simple linear interpolation.

In fact there is nothing else that DJI could have done because your drone does not know that WP1 and WP2 use the same POI - it does not know anything about POIs at all!!

POIs are a feature of the Litchi and VLM software only and are not transmitted to the drone. The drone only knows the altitude, heading and gimbal pitch at each waypoint and interpolates linearly between them.

In this case where two consecutive waypoints both use the same POI, I could, as you point out, do the interpolation differently so as to keep the camera pointing at POI2 for the duration of the leg - but since the whole point of VLM is to mimic the real flight there would be no point in doing that!

If you want to have the camera pointing continually at POI2 as you fly from WP1 to WP2, you will need to generate additional waypoints between them

I hope that helps.

Neil

You said it as I thought it- (and wrote previously)
We both explained similarly.

I figured that POI's are only a function of the software. (Not the drone and are used to just to calculate angle to send the drone)
… but the slow movement of the Gimble must be a function of the drone - because its clear that the drone continues turning even when its NOT in contact with the RC)
also... (If Not)
… but then Litchi - should be calculating CORRECTLY using the right Math.
ie ATAN function of distance and Height
not … Linear function of the angle.

See Handwritten Diagram (of altitude and angle between wp1 and wp2 looking at poi2)

on the left the angle is 2deg - Half Way its like 20 deg and … finally at Only 50m to go its an angle of 60 degrees- (at same alt of 100m all the way)
That's Logarithmic - and calculated as said before (prev Msg)
Litchi is calculating Linear angles from 0 to 90 as distance increases (Like halfway 45deg - not 20deg)

I also get what you said about your export as csv and mimicking what Litchi Actually does.
… in that you could do it more accurately but that would not mimick Litchi.

I Havent tested the actual Litchi mission as designed (as the wp's are too far away) but if ur sure that Litchi is miscalculating then is their problem - (Smile not yrs)

Thanks for looking into this - I knew you would quickly pick up the problem.

For more accuracy - please refer to the correct math for calculating angle on the Previous msg.

Thanks Neil.
Rob

 

[namirda]

You said it as I thought it-
We both explained similarly.

I figured that POI's are only a function of the software.
… but then Litchi - should be calculating CORRECTLY using the right Math.
ie ATAN function of distance and Height
not … Linear function of the angle.

I also get what you said about your export as csv and mimicking what Litchi Actually does.
… in that you could do it more accurately but that would not mimick Litchi.

I Havent tested the actual Litchi mission as designed (as the wp's are too far away) but if ur sure that Litchi is miscalculating then is their problem - (Smile not yrs)

Thanks for looking into this - I knew you would quickly pick up the problem.

For more accuracy - please refer to the correct math for calculating angle on the Previous msg.

Thanks Neil.
Rob

Hi Rob,

Litchi are not miscalculating anything really - my point is that they cannot do anything else because the drone does not understand POIs.

If you know that two waypoints use the same POI then, you could of course do the interpolation more cleverly so that the camera will always point at the POI as it flies from WP1 to WP2. However, Litchi need to upload the mission to the drone using DJI's API which doesn't know anything about POIs - so they have to keep it simple.

This is not anybody's mistake and it is not a bug in either Litchi, DJI or VLM - it is simply a limitation of DJI's API that does not use the concept of POIs

Perhaps we should call it an unimplemented feature !! Something for DJI to look into in a future update perhaps.

N

 

[Rockowe]

Hi Rob,

Litchi are not miscalculating anything really - my point is that they cannot do anything else because the drone does not understand POIs.

If you know that two waypoints use the same POI then, you could of course do the interpolation more cleverly so that the camera will always point at the POI as it flies from WP1 to WP2. However, Litchi need to upload the mission to the drone using DJI's API which doesn't know anything about POIs - so they have to keep it simple.

This is not anybody's mistake and it is not a bug in either Litchi, DJI or VLM - it is simply a limitation of DJI's API that does not use the concept of POIs

Perhaps we should call it an unimplemented feature !! Something for DJI to look into in a future update perhaps.

N

Mmmm I get what you saying -
ALSO iv noted that the drone gets AWFULLY confused when Out of Range from the RC - to do a Transition of gimbal Angle.
- So Then (at present) it would HAVE to be a Function of the WAYPOINT program (ie LITCHI) to calculate these better- and send the correct Math angles during the transition-

--- Please refer to the detail previously … ie
==============================

Litchi - should be calculating CORRECTLY using the right Math.
ie ATAN function of distance and Height
not … Linear function of the angle.

See Handwritten Diagram (of altitude and angle between wp1 and wp2 looking at poi2)
View attachment 56489
on the left the angle is 2deg - Half Way its like 20 deg and … finally at Only 50m to go its an angle of 60 degrees- (at same alt of 100m all the way)
That's Logarithmic - and calculated as said before (prev Msg)
Litchi is calculating Linear angles from 0 to 90 as distance increases (Like halfway 45deg - not 20deg)

==============================

Thanks for your Explanation.

 

[Rockowe]

In fact there is nothing else that DJI could have done because your drone does not know that WP1 and WP2 use the same POI - it does not know anything about POIs at all!!

… by the way thank you for your INSIGHTFUL explanation of this...