Antennatracking with opentx

[bobosch]

Is it possible to use the trainer port of the TX to control two 180° servos?
opentx could use the telemetry information to calculate the position of two 180° servos as shown in
http://fpv-community.de/showthread.php?157
http://vimeo.com/3991479
This would be an easy and cheap solution for fpv antenna tracking.

[thomas9x]

Cool idea. I think the required code space would be large, so this may be a challenge on a stock 9x. BTW, rather than use the trainer port, how about a second Rx that has the antenna tracker's pan & tilt servos on two unused channels, such as Ch7 & Ch8? That way there wouldn't be any wires from the 9x to the antenna tracker.

- Thomas

[kaos]

then put a 20mW vTx to relay the audeo/video signal to goggle. No wires either way. I hope Mike and Brent are reading this. Certainly the ARMS in sky9x or Taranis can handle this.

[mhkabir98]

I think we can add a second smaller atmega 328 board to add this support. Something like the audio mod. We could also add a GPS then to calculate current position and track antenna accordingly.

This combined with Mavlink would be awesome for autopilot use.

[pcwizzul]

This would be definitely great! I was gonna ask for this too. I'd like to have a camera recording my whole flight and a mechanical pan tilt like the one used for antenna tracking would be perfect.
Could we use the bluetooth streaming too? That way no channels would be needed but an external unit that controls two servos based on the data it receives via bluetooth.

[bobosch]

The intention was to keep it simple - yes it's the easiest way to use a second RX! So even a novice (like me ) could build this tracker.
Just hold the plane over the tracker and press a button to store the "home" coordinates.

An additional module would make it more complex but it could hold a GPS or bluetooth module... (hope it doesn't matter if the second GPS isn't placed on the tracker), but BT could interference with the TX.
If you need a microcontroller anyway, you could also put it together with a electronic compass into the tracker like the comercial trackers... just the question how to transmit the plane coordinates.

[bobosch]

Who could program the simple solution (without further hardware)?

[dvogonen]

This is a worthwhile project and a few people can do it if they want to. But it is not super simple since it would require updates to the mixes calculations.
I suggest that you make it into an official feature request via the openTX project page. I think that it is possible to report issues and make requests for everybody.

But the feature will probably not fit in a standard 9x/9xr. The program memory is almost depleted already for these radios. It is not possible to add more major features.
The function would be perfect for a Taranis, which has both the needed telemetry and enough memory for the feature.

[thomas9x]

Who could program the simple solution (without further hardware)?

Keep in mind that the "simple" solution is still a major project that would require significant effort. So given that the main code contributors are already being pulled in every direction, I think the antenna tracking feature would need a new OpenRC contributor. Or try bribing Mike, Bertrand, or Kilrah with fancy gifts.

The function would be perfect for a Taranis, which has both the needed telemetry and enough memory for the feature.

I completely agree. Besides the obvious pan/tilt code, there needs to be a menu interface for calibrating the mechanics. The calibration could be off-loaded to a PC app, but it would be more convenient if it was a Tx menu feature. Regardless of how all the little details are handled, I would be amazed if someone was able to make a usable version fit in a stock 9x. But the Taranis would have *all* the resources already built into its stock hardware.

Even though interest in Antenna Tracking has cooled down a lot in the last year, it's still a awesome idea. And using a second R/C Rx as a "wireless" link to the tracker's pan/tilt servos is a nifty concept too.

- Thomas

[Kilrah]

Yep, tracking built in the radio would be great - heck, it might even become simple enough for me to actually use an antenna tracker one day

The Taranis' 16 channels are also a benefit, anybody will nearly always be able to spare 2 for that, bind a 2nd receiver and put it on a tripod with just 2 servos without requiring a separate RF link.

As to development yes there is a lot of cool things we can do and that are requested. To save work we're trying to compile them and avoid coming up with solutions that are dedicated to just one task like a tracking-specific menu would be. For example some people would want to use math functions to drive special kinds of models like ship turrets, which basically isn't that much different in terms of requirements from antenna tracking. If we could come up with a solution that works for both, and other potential uses we'd be happy.

The interactive templates/scripting system proposition I did some time ago (here) would do it. It's basically a plugin system, with a settings page for each special function that allows setting parameters and triggering actions without adding anything in the code. It would work just as well for those special uses than for common models and thus benefit everybody, be highly flexible yet simpler to program than having to dig into the openTx code, etc. Plus the time to implement it would probably not exceed by much that for a single dedicated function.

[bobosch]

Yes, with "simple" I've thought without additional hardware in the TX ... You've wrote it already, it would be a challenge in stock 9x.

Feature request - should I write a summary here in this thread or add a entry in the issue tracker as "Enhancement" ?

[jhsa]

Now you just need video signal on the radio's display and a OSD displaying the telemetry over the video

Or at least the radio generating the OSD signal to be mixed to the video signal to your goggles.

João

[bobosch]

I've tested this two-receiver-idea with the FrSky D.. receivers. This works not, because telemetry (at least the voltage information) always shown from the nearest receiver...
I've read that it's possible to bind current V.. receivers (for use in the tracker without telemetry) also in D-mode. I will try this as soon as I have one

[bobosch]

Or at least the radio generating the OSD signal to be mixed to the video signal to your goggles.

I love this idea ... but I already have an OSD

[Kilrah]

I've read that it's possible to bind current V.. receivers (for use in the tracker without telemetry) also in D-mode. I will try this as soon as I have one

Yes, you would want to use either one D and one V-II receiver, or 2 X receivers with one of them having telemtry disabled.

[bobosch]

Back to the original idea:
- Add a custom function to store current coordinates as tracker position
- Calculate servo position for the two 180° servos with the coordinates to track plane
- Add the result as new sources in the mixer (so calibration and channel assignment is possible with mixer functions)

Will this fit into a stock 9x? Anyone who like to try coding this?
As I won't need to buy a standalone tracker, I have some resources for donating

[tilmanb]

Bertrand let's move this to a separate thread.
Bobosch, it would be possible but unlikely, non specific enough to justify the space used in the default build.
Code wins. Give out a shot

[bertrand35]

Back to the original idea:
- Add a custom function to store current coordinates as tracker position
- Calculate servo position for the two 180° servos with the coordinates to track plane
- Add the result as new sources in the mixer (so calibration and channel assignment is possible with mixer functions)

Will this fit into a stock 9x? Anyone who like to try coding this?
As I won't need to buy a standalone tracker, I have some resources for donating

Understood

1st step - we add the lua interpreter into Taranis / sky9x

2nd step - we work on the maths in lua for that

3rd step - we write the algorithm in fixed point, we check it doesn't use too many resources on the 64A

4rd step - we add it to the code with an option, with 2 new entries in the mixer sources list

[bobosch]

Bobosch, it would be possible but unlikely, non specific enough to justify the space used in the default build.

Yes, it's unlikely, but as module like the voice option it wouldn't waste space in the default build.

[bertrand35]

1st step is done. Here is the start point for 2nd step. I will let the mathematicians work...

[kaos]

Fantastic to see this project moving forward in such great paste. I hope Mike will port this to Esky9x. Don't have a Taranis but have a Sky9x with 16 ch to use.
What is the 'bribe' Thomas was talking about? any hint?
Donate implies to helpless people. Bribe implies to able, people with power that sounds right. May be we should add a 'Bribe' button beside 'Donate'.

[bertrand35]

So ... no trigonometry expert here?

[MikeB]

Not quite sure what the question is!
Is it we have a GPS position of the Tx, and a GPS position of the Rx (in the air) and you want to know the polar coordinates of the Rx from the Tx referenced to some reference direction (N-S?)?

Mike.

[bertrand35]

Yes we will need to determine the angles for the 2 servos of the antenna which follows the plane. But I thought that you (and me) have the plate quite full, so I was asking for help of somebody else wanting to give it a try

[Rob Thomson]

I would... but me and maths.... it would be carnage.

[kaos]

Remind the mathematician there are 360 degree servo: http://www.alibaba.com/showroom/360-degree-servo.html
or 1260 degree servos : http://www.robotshop.com/hitec-hs785hb-servo-motor.html

[thomas9x]

The "360" robot servos will not work in a tradition pan/tilt system. They are continuous rotation (not proportional) and there is no way to control their exact position. The 1260 degree winch servo can be used if it has reasonably accurate proportional control. But cable management needs extra attention to prevent the wiring from wrapping around and around. Even with a winch servo, the software still needs to keep track of the number of full rotations so it knows when it has hit the rotation limit. Then the "unwinding" process needs to occur.

The most common solution is to use a standard hi-torque servo with gearing on the pan axis. But a hefty servo that can achieve about 300 degrees of rotation is fine and most FPV antennas will achieve good RF beamwidth coverage with this amount of travel. As usual, the software needs to know when a rotation limit has occurred, then it must unwind the servo. FWIW, the travel limit & servo unwinding related parameters are configured in the setup procedure of the popular antenna trackers.

- Thomas

[kaos]

I was surprised there is no 360 degree proportional servo available. and winch servo certainly is cumbersome.
After looking into the tracking system a little more. I am wondering when using only one 270-300 degree pan servo almost prevent someone flying in a circular pattern. What if we use two 180-270 degree pan/tilt system and combine the diversity antenna protocol?
use two sets of Tx/Rx ant. each one is responsible for 180 degree of tracking and with diversity circuitry only the 'best' signal will come across.
The program needs to track where the servo of the best signal is (based on the GPS position/direction) and the other one will be moved to the nearest 0 or 180 degree position to pick up the signal when one ant reaching the end point while the other helical will be ready to pick up after 180 degree. so the program needs to know the servo position of one ant and the other one is always ready do pick it up when one servo is reaching 0 or 180 degree. Basically the two pan servos will move at the same direction, one is moving from 0-180 while the other is moving from 360 to 180 and vise versa. while both tilt servo angle is going to be the same from GPS data.

Hmm, this can even be used for the Tx/Rx ant onboard the plane.

May be this is off topic. move to a new thread if moderator think it is appropriate.

[jhsa]

In my opinion and as someone said before, the best option might be using some gears between the antenna base and a normal servo. this would allow the antenna to go 360º.

João

[kaos]

true, but if there is a gear ratio that would decrease the accuracy/sensitivity of the tracking and it does not solve the problem that you can not fly in a circular pattern, the pan servo has to turn almost 360 degree backward when the plane past beyond 360 degree. If you are flying back and forth at the 360 degree area, it is going to be hectic.
I don't know any one would fly in circular pattern on purpose or not. But in theory that is a limitation.
with 2 sets of 180 degree tracking with diversity, it would be seamless however you fly.
another idea is, the tracking system I read about all need to have the tracking system set up pretty accurately of their direction (for GPS reference). I also wonder how this can be streamlined with a magneto sensor like the one used in a multicopter control board. If this can be integrated, you just throw the tracking rack (like tripod) on the ground, turn it on, it will set up itself. with the tracking program Bert is working on, it would make the tracking system a lot easier to use and practical.
I think a lot of people don't like using a tracking system is because it is cumbersome to use. But if it is pretty much one switch operation, a lot of people would have used it.

I don't mind to be the GP to build a system like that if I can get some technical help here.