Is CCC more precise than UAV Toolbox?

[Skeets]

Hi folks!
I was wondering if setting up CCC waypoints via the ST-16 would be more precise than using a waypoint planner such as UAV Toolbox. For example, if I were to save a CCC point and then run the CCC mission, will the H go EXACTLY where I saved that point? UAV Toolbox is great, but the precise altitude and/or coordinates may be slightly off due to the maps (Google, Bing) inaccuracies.

Thanks!
Ed

 

[PatR]

Your use of the word "exactly" suggests you are looking for performance that does not exist at this level of aerial robotics. My concern with offering any kind of response is that you intend to use that information to build very low level, minimum separation missions that would have high probability of failure, for which the H would be blamed. None of the flight controllers used in consumer drones provide the level of precision you appear to want. You need to step up to custom or enterprise level equipment to obtain it.

 

[Steve Carr]

The most precise method is to fly the route to set up the waypoints. I can't see any difference using the ST16 or the Toolbox. They would be using the same maps. The altitude is going to be the most troublesome because it is relying on the barometric altimeter. You want to give yourself lots of room. Also remember the route will be flown in a curve to the waypoints.

 

[Skeets]

Your use of the word "exactly" suggests you are looking for performance that does not exist at this level of aerial robotics. My concern with offering any kind of response is that you intend to use that information to build very low level, minimum separation missions that would have high probability of failure, for which the H would be blamed. None of the flight controllers used in consumer drones provide the level of precision you appear to want. You need to step up to custom or enterprise level equipment to obtain it.

See? That is the problem with these kinds of forums. Too many assumptions! I have absolutely no desire or intention on doing anything remotely (pardon the pun ) risky or dangerous to people, animals, plants, vehicles, buildings or my aircraft (and please don't assume that because I didn't list it, it is open season). I may, however, at times require more precision that I thought maybe CCC would be a better choice. My question has nothing to do with low level minimum separation missions so you need to get off your high horse and not tell me to step up to custom. I've been in the RC game for over 15 years and I know what is acceptable custom of which this is not even about. I take full culpability in what I do and would absolutely not hold Yuneec, or anyone for that matter, accountable for my own actions.

 

[Skeets]

I can't see any difference using the ST16 or the Toolbox. They would be using the same maps.

Are you sure? I thought that when using CCC, it recorded the lat/long and altitude of the H at the moment you press the +. I do not see any maps being used when doing it that way.
Thanks

 

[PatR]

Understand, but I've been a long time involved with this stuff and have seen a great many questions start out relatively innocent that shortly turn into a product blame game. Worse, far too many accept any answer to a performance question, especially those that provide some level of approval for something they want to do, as a hard truth. After doing so they go out and put the answer to actual use without progressive testing. From there things often go down hill fast. The word exactly, in caps, well qualified my earlier response. That position has not changed.

I will go as far as saying the H is far more accurate in CCC position repetition, with both aircraft and camera, than similar products of other brands. Tuna's software changes how you can input those positions but does not appear to alter how they are executed. However, the H is not a repeatable precision instrument.

 

[Skeets]

Understand, but I've been a long time involved with this stuff and have seen a great many questions start out relatively innocent that shortly turn into a product blame game. Worse, far too many accept any answer to a performance question, especially those that provide some level of approval for something they want to do, as a hard truth. After doing so they go out and put the answer to actual use without progressive testing. From there things often go down hill fast.

I understand, but not everyone asks questions with the intent on doing something unsafe. We both know that, unfortunately, there are folks out there who are going to give the rest of us a bad name, but it is not for you, me, or anyone else to determine what questions or advice that is asked is for the sake of doing something unsafe. Unless a question that is asked that is so blatantly obviously unsafe ("how low can I fly over an airport?"), then perhaps the benefit of the doubt should be

 

[PatR]

Point taken. Unfortunately we have a risk management condition to consider; lacking qualifying background the worst case scenario has to be assigned the pinnacle position when generating a response. In this case, I think I understand the way you often think but I don't know I understand the way you often think, and discounting that I need to consider how others might view or make use of an answer. The latter can be much more important than the former.

For your original question, perhaps some experimentation through multiple flights using Tuna's App would be beneficial in developing statistics relative to positional repeatability.

 

[Steve Carr]

Are you sure? I thought that when using CCC, it recorded the lat/long and altitude of the H at the moment you press the +. I do not see any maps being used when doing it that way.
Thanks

Oops....I was thinking of the time when the ST16 did display a map. That is no longer the case. I think we can agree the CCC method is going to be most accurate and repeatable. But I know many people have found a repeat flight often is slightly different each time it's flown. Again, it's good to build in a lot of clearance when planning the route.

 

[Tuna]

Great question!

Whilst it's possible that Google/Bing maps have some inaccuracies, they are on the whole pretty small as the map data has to be made to line up across the whole globe. Some basic tests on the web suggest that typically you're looking at errors in the range of inches rather than yards.

When you're flying, relying on single frequency GPS (which I think the Typhoon does), there are also inaccuracies. The US Gov suggest you'll get accuracies within six feet 95% of the time. In practise I think the sort of conditions we usually fly in produce better results, but you should not rely on positioning being absolutely reproducible between flights.

The other really important thing to realise is that when you fly a CCC route, the Typhoon does not fly it as you originally recorded it. If you fly a perfect square and put a waypoint at each corner - the Typhoon will fly a circle when you replay that route. I can't emphasise that enough - the Typhoon flies a curved route through the waypoints that is completely unrelated to the route you flew to record them. It attempts to 'smooth out' the route to give you a clean curved flight path.

With the above example of a square flight path being turned circular - a 100 foot square will see the Typhoon deviate by over 20 feet between waypoints.

UAV Toolbox does its best to estimate that flight path. So when you put down waypoints you can see what happens between them. It's not perfect as I had to reverse engineer what the Typhoon does when it flies a CCC route, but the results appear to be reasonably accurate.

So there's your answer. If you want to fly to a single point, CCC and UAV Toolbox are both probably about equally accurate - but you should not rely on results closer than 6 feet of your original point. In practise you should get better than that, but I wouldn't plan points near power lines or buildings. If you want to fly a route accurately, UAV Toolbox should be much better than pre-flying a CCC route as it estimates the curve.

In either case, always be prepared to abort a cable cam mission if it looks like the Typhoon is heading in the wrong direction.

 

[Bartcephus]

Ah ha.... that explains it.... it uses the barometric alt, not the GPS alt.... so that's why it was 12 feet too low and ended up in a tree at 66 feet up. Argh.... Powerline right by the tree. I got lucky, the power company came out and rescued my drone rather than shut off the power to let me shimmy up the tree... (good thing too, I am NOT a monkey or squirrel.

I was wondering why it did that.... I had 20 GPS Sat's locked in on both units.

Best Regards,

Bartcephus

 

[PatR]

Tuna,

Most advanced autopilots fly a way point route making directional changes at a tangent to a way point, not directly over one. Doing that moderates abrupt directional changes. With a few more advanced auto pilots the user can adjust the distance from a way point the auto pilot will initiate a directional change.

 

[vr-pilot]

Tuna,

Most advanced autopilots fly a way point route making directional changes at a tangent to a way point, not directly over one. Doing that moderates abrupt directional changes. With a few more advanced auto pilots the user can adjust the distance from a way point the auto pilot will initiate a directional change.

UAV toolbox is a more advanced autopilot tool, as it is able to define the path curvature at waypoints and depicts it quite accurately on the map.
My flight tests or usage of UAV toolbox shows, that it is as accurate as GPS allows. I would at least expect a 5 meters GPS accuracy radius. That results in 10 meters maximum variation (circle diameter) when repeating a CCC mission.
I bought UAV toolbox some time ago and I am really amazed by all its features like equipment list, mission planing and mission analyzation (e.g. video and map view synchronized).

 

[Air pics]

I was hoping to stitch parallel flight paths together to get an over all look at a large area. Looks like I am out of luck with my H.

 

[vr-pilot]

I was hoping to stitch parallel flight paths together to get an over all look at a large area. Looks like I am out of luck with my H.

Maybe you will find this tool useful for area mapping (provides mapping grid):
Flight Planner H - Waypoints Creator for Yuneec Typhoon H
I just created 9 CCC missions with it today, as it is the only editor I know that is able to create those scan type patterns.

 

[Tuna]

Tuna,

Most advanced autopilots fly a way point route making directional changes at a tangent to a way point, not directly over one. Doing that moderates abrupt directional changes. With a few more advanced auto pilots the user can adjust the distance from a way point the auto pilot will initiate a directional change.

Exactly. The challenge with the Typhoon flight controller is that the parameters used are not published anywhere, so had to be estimated by flying quite a few CCC missions to test the different behaviours. In the case of the Typhoon, the change is continuous and the distance is weighted according to the distance between waypoints. So if you want to do a tight turn, put two waypoints close together to tighten up the curve.

 

[Tuna]

Maybe you will find this tool useful for area mapping (provides mapping grid):
Flight Planner H - Waypoints Creator for Yuneec Typhoon H
I just created 9 CCC missions with it today, as it is the only editor I know that is able to create those scan type patterns.

Indeed, until I get scan working in UAV Toolbox, this is a good option. Though I've found that you can manually draw out scan flight plans for smaller areas with no problem. If you're using a stitching program to create orthomosaics, you don't need a perfect straight flight, just sufficient overlap between photos. I got very good results last week scanning a 20 acre site that way and it wouldn't be a problem to scale it up further.

 

[360Media]

Hey Skeets. I use CCC fairly often in my real estate photography business. I love it. One job I used it to fly from lake to just over power lines, then through a complex maze of trees and landscape on the way to the house. Altitude fairly precise and repeatable, and coordinates very precise and repeatable. It just takes time to set up the CCC, but I found it works perfectly.

Hi folks!
I was wondering if setting up CCC waypoints via the ST-16 would be more precise than using a waypoint planner such as UAV Toolbox. For example, if I were to save a CCC point and then run the CCC mission, will the H go EXACTLY where I saved that point? UAV Toolbox is great, but the precise altitude and/or coordinates may be slightly off due to the maps (Google, Bing) inaccuracies.

Thanks!
Ed

 

[vr-pilot]

Indeed, until I get scan working in UAV Toolbox, this is a good option. Though I've found that you can manually draw out scan flight plans for smaller areas with no problem. If you're using a stitching program to create orthomosaics, you don't need a perfect straight flight, just sufficient overlap between photos. I got very good results last week scanning a 20 acre site that way and it wouldn't be a problem to scale it up further.

Your program is much more sophisticated and I bought it a while ago.
The only thing that I found "missing" was this automated scan field generation. In fact your program allows multiple waypoint selections and moves etc. so that a scan pattern can be made "manually" quite fast.
When I have you just "on the line":
Your program offers the selection of different cameras for the FOV depiction. Very good idea. I can select between CGO3+ (Pro) and two types of lens mods. There are two CGO3+ kit lenses sold by Yuneec: the 98° FOV (fomerly US only) and 115° FOV (formerly EU only). As it seems now all new "CGO3+ (Pro)"s come with 98° FOV. Would be fine to be able to selcect between them.
And because I am an owner of a CGO-ET (just since 12 days), it would be fine to select its "footprint", too. The FOVs of the two built-in cams (Low-Light 16:9 and Thermal 4:3) are different though. That directly influences the path width of the scan pattern and its overlapping areas respectively...

Thanks for the great program. Looking forward to use it with the H520!

 

[Tuna]

Your program is much more sophisticated and I bought it a while ago.
The only thing that I found "missing" was this automated scan field generation. In fact your program allows multiple waypoint selections and moves etc. so that a scan pattern can be made "manually" quite fast.
When I have you just "on the line":
Your program offers the selection of different cameras for the FOV depiction. Very good idea. I can select between CGO3+ (Pro) and two types of lens mods. There are two CGO3+ kit lenses sold by Yuneec: the 98° FOV (fomerly US only) and 115° FOV (formerly EU only). As it seems now all new "CGO3+ (Pro)"s come with 98° FOV. Would be fine to be able to selcect between them.
And because I am an owner of a CGO-ET (just since 12 days), it would be fine to select its "footprint", too. The FOVs of the two built-in cams (Low-Light 16:9 and Thermal 4:3) are different though. That directly influences the path width of the scan pattern and its overlapping areas respectively...

Thanks for the great program. Looking forward to use it with the H520!

Drop me a message and I'll explain what measurements I need from the cameras/lenses to add them to the list.