Jump to content
Mr.RC-Cam

Artificial Horizon

Recommended Posts

At the old MSN RC-CAM site, there had been several discussion about adding an artificial horizon indicator (AHI) to our video camera equipped models. There was another short thread about it here a few months ago. Is this something that folks are still interested in? Has anyone designed anything that works well?

If not, what would be the ideal product features for something that did this? Other than free. wink.gif

Share this post


Link to post
Share on other sites

Hi Mr. RC-CAM,

This idea is great! I am all for it. Data modem and ground station projects that were discussed earlier on this forum are related to artificial horizon display. MikeP has opened dedicated forum to develop ground station application capable of artificial horizon display as well as multitude of telemetry data. I am currently trying to test fly data modem that will send telemetry including attitude information obtained from co-pilot sensor. In my view the next step is to finalize the started projects and to show in practice that suggested designs are capable of AHD at refresh rates that are usefull to pilots. But I bet you must be thinking of something else ;)

What do you have in mind?

Regards,

Val.

Share this post


Link to post
Share on other sites

The artificial horizon is a display device that tells you what the aircraft's attitude is. All full size pilots rely on them.

Cyber-Flyer, I was interested in taking the Co-Pilot type IR technology and using it to create an artificial horizon. The IR method would be ideal for us. It sounds like you are into this too. If you have already done this then there is no reason for another cook to enter the kitchen.

My implementation was going to be determined once I understood what others wanted/needed. The display could be nearly anything. Perhaps an LED array in camera view, RS-232 output scaled as needed, voice output, or perhaps an integrated OSD. My thoughts are that it will need to be cheap, so the LED or RS-232 output would be the likely solution.

Anyway, if there is any interest then I will listen to what folks want and see if it makes sense to go further. But again, if you are deep into the project already then I don't think I would be too excited to start another me-too design.

Share this post


Link to post
Share on other sites

When you are flying at low altitude in familiar territory, the need for it is minimal. However, at high altitude or in poor visibility, the artificial horizon offers an accurate sense of what the model is doing. This is also true for full scale aircraft.

Maybe others will chime in and explain why they needed it. There were several folks that were trying to create this sort of thing for there own unique R/C video application.

Share this post


Link to post
Share on other sites

The FMA copilot generates the info needed for an attitude indicator. It is reasonable in cost and availability and many people already have it. It would be quite useful to learn how to tap into it and get the signals out. My thought is that you would "reach" a greater user audience by first going for the integrated video overlay. And if the overlay could display attitude and the typical GPS data then you have a complete flyable/navigable system for an RC pilot.

RS232 would require additional downlink parts/systems on the ground. Data downlinking beyond video imagery, although quite useful, is not required for the simple 1st person view type of experience, which is what I think would engage most hobbyists. The in-view leds might be ok, but is optically more finicky than it might first appear due to having to work with a camera that mostly wants to look unobstructively outside of the ship at focal infinity. I would expect such a contrivance to generate a fuzzy, overly course kind of display.

Although not nearly as elegant as the video overlayapproach, probably the least expensive attitude indication might be to use the audio channel that many video xmtrs already have to send simple pitch/volume modulated sound to a ground speaker/headset. No further equipment required except for a trained ear.

Share this post


Link to post
Share on other sites

....And yes, an attitude indicator would be extremely useful when flying at higher altitudes. Here are some ideas.

I have planes that can be very difficult to judge from the ground whether flying up-side up or up-side down. (A flying wing is one.) A wind gust or mis-control can easily upset a plane and the attitude indicator can help the ground pilot with prompt recovery. True, the co-pilot should keep plane level, but then you are not the one flying..no sport in that B) . Furthermore, if plane is climbing or descending (as usually happens after an inversion) , the camera may or may not show the horizon. The attitude indicator is the only way (again, if you are doing the controlling) to quickly compensate for unwanted, energy dissipating, rolls. And if you accidentally get your plane into a bad visual situation ( haze, thermal into a cloud, low level light, or your view of the plane is suddenly obstructed), the attitude indicator would be helpful. One more idea...if you have an attitude indicator, you can start to learn to control aircraft pitch vs speed the way full scale pilots do and this can be fun. For example, for a given power level you can hold a certain pitch on a high performance climbout and reasonbly expect not to stall.

Well, a little of the above depends on what your missions or desires are, but they point to the utility of the thing.

Share this post


Link to post
Share on other sites

When I read the first post, here are the thoughts that popped into my head.

- I have a few planes that require some down elevator when in a power climb.... to little elevator and the nose balloons up and the plane stalls or rolls off to the side. Too much elevator and I waste precious energy in a highspeed, but shallow, climb. A pitch and roll sensor would help me maintain an optimal flight path when the horizon is not visible.

In terms of the implementation, my preference is to have all of the logic on the airframe. In otherwords, I don't want an extra ground component that captures some signal from the plane and then overlays it on the video.

Next, I think that an overlay would be the slickest implementation. Something as simple as a bright dot on the right hand side for pitch and a bright dot on the bottom for roll.

The next best thing would be LED's placed somewhere in front of the camera. They don't even have to be perfectly in focus, they just have to be visible. Perhaps a clever way to do this is with those nice compact LED voltage watch thingies. Then your PIC doesn't even have to drive individual LED's, it just has to generate two voltage outputs. The only downside I see here is most voltage watch units use color coded LED's, and I'd think you would want just one color of LED's for this project.

And finally, I would want it to be insanely compact to fly it on even my smallest electric airplanes.

That's my wish-list :)

Regards,

Bill

Share this post


Link to post
Share on other sites
Cyber-Flyer, I was interested in taking the Co-Pilot type IR technology and using it to create an artificial horizon. The IR method would be ideal for us. It sounds like you are into this too. If you have already done this then there is no reason for another cook to enter the kitchen.

By all means Mr. RC-CAM go for it. There must be several ways to skin this cat.

This project is complex enough that it can stamble at ever increasing amount of underwater stones and never be completed. I am all for yet another solution, if it is available. Hey, I even tried very "simple" AHI some time ago:

http://www.cyber-flyer.com/2001/flight092301.html

I do think that telemetry + ground station is the most exciting way to go. But I also like sound encoding of attitude idea for its simplicity.

Currently, I am trying to cram data modem and GPS navigation in a single chip system. (Bill, I am working hard on your dream). So my project may take long time to complete. From where I stand, it makes sense to join your talents with MikeP and work along to build telemetry board that will mate with his groundstation. He has put telemetry specs on his forum:

https://sourceforge.net/projects/rcpilot/

and got few people writing ground station app.

Regards,

Val.

Share this post


Link to post
Share on other sites

Yb2normal, your overlay comments are what I had in mind if I used the OSD method. It would use a little transparent glyph on the left side for pitch, and another on the bottom edge of the video frame for roll. They would move to indicate the attitude position. With luck I would also allow it to display as a moving horizon line, like full scale displays often use.

It sounds like the OSD is big on the list. The full project is complicated enough that I would probably have to offer a PCB set and maybe even a parts kit (some of the parts would probably be difficult for the average hobbyist to buy). Overall, the cost to develop the project is not looking good to me. I'm still a bit bashful after the bust from the video switcher project.

My goal is to create something that others could replicate at low cost. The audience for this sort of device is already small, so my thoughts are that it has to be under $150 to get any real attention.

So, I will need to ponder on the project a bit more. In the meantime, please keep the "must-have" feature suggestions coming.

Share this post


Link to post
Share on other sites

I'm still looking for ways to keep cost down. For the OSD, using the commercial video overlay boards (kit or ready-to-use) will burden the project cost a bit. They all seem to start at nearly $100. Also, they all seem to be a bit larger than necessary.

I would like to use the STV5730A OSD chip, but it is a bear to buy. So, I am thinking of creating my own genlocked video overlay board for the moving cursor glyphs. The horizon graphics would be two small tick marks that moved along the left (pitch) and bottom (roll) edges of the screen. There wouldn't be any text capabilities, so adding nifty features like a call sign pop-up would not be available.

Cost would be much lower to do it this way. Would this solution be popular with you folks?

Share this post


Link to post
Share on other sites

Yeah, I have some of the bob 3 boards and I really wish they would use something other then the old 30 pin DIMM boards. The board could be so much smaller then it is and still get the job done.

Matt

Share this post


Link to post
Share on other sites
The horizon graphics would be two small tick marks that moved along the left (pitch) and bottom (roll) edges of the screen.

I'd love to have something like this as addition to TL100 screen!

Done! Put me as a beta tester, no put me as an alfa tester.

Share this post


Link to post
Share on other sites

My vote is for the simple tick marks on the borders of the screen. And if the circuit fits on my thumbnail, all the better ;)

Bill

Share this post


Link to post
Share on other sites

Well I took a break from the fun for a while to do some of the necessary stuff in life. But I started working on the ground station again today.

Basically I added support for APRS from the TinyTrak. I know this does not help with an artificial horizon, but it does make it easy for someone that might want to use it try it out.

Basically the same old setup of TT to TX. RX to AGWPE.

It's not my prefered method for receiving telemetry (slow and not much info) but I'm hoping it will help the project gain some acceptance to encourage development.

Cheers,

Mike

Share this post


Link to post
Share on other sites

I too would be enthusiastic with the edge glyph attitude display. However, it might be more intuitive, if not traditional, if the two glyphs were on opposite sides of the screen. In that way a line (imaginary or actual) could represent pitch & roll by whether it was high/low and flat/diagonal. And just because it is easy at this stage to wish for the moon, how about remotely controllable on/off glyphs to facilitate recording clear video?

Share this post


Link to post
Share on other sites

Sorry for continuuing, but a question I have is how to handle an over range situation? Imagine the horizon falling below the field of view of the camera, would the glyphs simply be pinned down? I guess I'm thinking that if they had exponential travel, they could still be yielding some info when over ranging? But, perhaps my head is in gimbal lock and I'm way off base here?

Share this post


Link to post
Share on other sites

RadioFlyer, I like your linked glyph idea. But, I am not sure I can do it. My low cost OSD idea is not expected to have the available CPU overhead needed to do it that way (jitter free video is a tough on a little microprocessor).

Over-ranging is handled with data point limiting (gyths are pinned at end of screen travel). I expect to scale the movement so that it falls within a useful range; severe pitch and roll movement will hit the travel limits. My proposed 2-glyph method works well with this since pinned data does not visually limit the other axis' movement.

The good news is that I am very confident that I can utilize an existing Co-Pilot (or just its IR head) to obtain the horizon data. It's the video overlay that is going to make me work a bit since I want to avoid a commercial board. My goal is to create a custom solution that is small enough for convenient use on a park flyer.

Everyone, thanks for the feedback and encouragement. The project is serious enough that it will be awhile (weeks) before anything can be shown. If an alpha model is available then Cyber-Flyer is first on the list! ;)

Share this post


Link to post
Share on other sites

Hi all.

I hav'nt been on this forum long ,so please forgive me if I'm stating the obvious.You guy's must have considered using accelerometers such as the Analog Devices ADXL 202 for your HUD project.It to me is a viable alternative to the FMA autopilot,which incedently I also use,to give a usefull output on the two axis required.

Cheers Johnstorm.

Share this post


Link to post
Share on other sites

Unfortunately the ADXL devices are of course sensitive to both static and dynamic acceleration. This will work let's say if your aircraft is flying absolutely stable, but not when you're moving around any of its axes.

If you have one of these "motion sensor" game pads (There is an ADXL202 inside), try to lay it flat and to hit it. You'll see that the cross goes to the edge even if the angle doesn't change. The other solution would be the gyro.

Share this post


Link to post
Share on other sites

And of course there is always the drift problem.

--------------------

Mr. RC-CAM. Would you be willing to share the FMA part with me.

I still plan on going the Ground Station route simply because the kind of data I want to collect is too abundant to be transmitted over video.

But I would really appreciate the schematic and theory behind using the FMA to calculate pitch and roll.

Regards,

Mike

Edited by mikep

Share this post


Link to post
Share on other sites

<_< Thanks guy's for the replies regard accelerometers.

Please indulge my curiosity,My understanding of this is much like my Skiing,"I try my best but often end up on my ****".

I can see how dynamic forces can effect the pitch of the vehicle,forward speed acceleration etc, but how does it come into play regarding roll?

My guess would also be that forward speed etc can be compensated for in a software 'offset'.Thanks for the indulgence.

Cheers Johnstorm

Share this post


Link to post
Share on other sites

Basically when you are making a perfectly coordinated turn, the accelerometer should be reading no tilt. But let's put that aside, they tend to drift which is one major problem. So what I've seen done is a combination of gyros and accelerometers being filtered through a Kalman filter to get reasonable results.

A very good example of this is the Autopilot project.

Take a look at http://rotomotion.com/prd_IMP.html for some kits.

Mike

Edited by mikep

Share this post


Link to post
Share on other sites
... but how does it come into play regarding roll?

Dynamic changes to ANY axis just creates G-Forces in that plane. So, upon an aircraft bank/roll, the related ADXL sensor sees G-Force (acceleration), along with the model's wing angle (static acceleration). It cannot discern between the two and they get lumped togther.

As an experiment, connect up your ADXL and measure it while you tilt it (static acceleration). You will see the voltage move with the angle, just as you expect. Then, hold it level (or any angle you wish), but move it about in free space (G-Force measurement). This simulates a model moving around. You will quickly see the trouble the movement causes -- it can overwhelm the static acceleration data.

It would take much more than an fixed offset term to subtract out the unwanted forces (the G-Force due to movement is not readily predictable). It would be very complicated to do so.

Would you be willing to share the FMA part with me.

It is simple a analog voltage from the IR module (one for each axis). The voltage increases or decreases, depending on the difference from the horizon. You could determine absolute position if you don't mind the overhead of the math (it would be easier to just xmit the two voltages to your groundstation and do the math there). However, I am going to create a simple relative readout. It will be like a carpenter's level that has a damped bubble.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×