Autopilot selection for electric UAV

[2bent]

Hi all-

I’m looking to set up an autopilot for a project I’m helping with over the summer. We’ve been looking around for autopilots for a 68” electric airplane, but aren’t sure exactly what is possible with what’s available. Ideally we’d be able to preprogram a route and altitude to fly, then launch and climb manually, switching to autonomous waypoint tracking and altitude control after climbout. We’d also like to have real-time GPS position reports transmitted to a ground station (laptop). It would be fantastic to be able to switch between autonomous and manual control whenever we need to, as well as update our waypoints on the fly.

We’ve been looking at the U-nav Picopilot NAT system, which looks great in that it’s pretty simple to install, and it can downlink GPS data to Garmin mapsouce in real-time through a radio modem. I gather that waypoint data for the Picopilot is entered on the ground before launch, but I’m not sure whether it can also receive the same waypoint data inflight through the radio modem (the U-nav 3400 is definitely able to, but that is just over our budget).

The RCAP2 system also looked intriguing, but it seems like a more involved installation process.

If anyone has any other suggestions or recommendations for systems to consider, as well as any other thoughts on data uplink or the whole project that’d be great!

BTW, last summer we built a similar airplane, but just used an FMA co-pilot to help level the plane (conventional planform, pusher prop behind the wing on a pylon), and used an Eagletree Seagull system for telemetry. I’ve been flying electric R/C for years, but this is my first experience with autopilots, so bear with me...

Thanks for the help,

-Ben

[icebear]

I gather that waypoint data for the Picopilot is entered on the ground before launch, but I’m not sure whether it can also receive the same waypoint data inflight through the radio modem (the U-nav 3400 is definitely able to, but that is just over our budget).

Yes, WP data is entered via a PC on the ground in the Picopilot. It is a very good autopilot and has the advantage of including wing-levelling so you wouldn't need a Co-pilot. It does not allow you to program altitude however. It keeps whatever alt you are at when you enable the autopilot. You can however turn the autopilot on/off in the air to change altitude manually.

The RCAP2 is a more basic autopilot and I would say that the features and tracking capabilities of the Picopilot is definitely more advanced.

Keep us posted and good luck!

/Icebear

[2bent]

We have decided on the Picopilot-- should arrive tomorrow or monday! A review of your amazing lighthouse mission definately helped!

We haven't yet made a decision on telemetry-- during tests we are using an Eagletree Seagull system for live GPS tracking, but would like to eventually use a radio modem (probably 9xtreme) to transmit the data straight from the Picopilot to the ground station, just as U-nav mentions. We're hopeful that it might be possible to upload new waypoint tracks in flight, which is one of the drivers for using a bi-directional modem instead of the Eagletree system in the final UAV.

I'll keep this updated as things progress...

-Ben

[icebear]

Ben,

Thanks for the comment on my lighthouse mission!

Interesting to see if you can manage to upload WP's in flight somehow.

I am looking forward to hearing more about your project!

/icebear

[2bent]

The Picopilot-NAT came in yesterday, and was installed on a brushless slowstick just to test it out. We ran into the same problem you had with the Picopilot not responding to our disable command (we are using an FS-8 co-pilot rx), where it did drive the servos, but wouldn't respond to the transmitter, or shut off the green LED. A set of servo buffers/boosters are on the way from FMA.

We did test it out in the car anyways, by setting a series of waypoints a mile or so apart along a curved road, and watching the rudder deflect as we drove. That was a positive test, although we of course won't be able to fly until the servo boosters are in.

About updating waypoints in flight-- to program waypoints into Picopilot, you simply unplug the GPS connector and plug in the serial cable to the same connector. We were thinking that some sort of switch (mechanical or electronic) could be installed to do this remotely, as the modem appears to go both ways, not just airborne to ground. Hopefully this is a realistic ambition, and hopefully the diagram makes more sense :)

-Ben

[icebear]

Thanks for the explanation about updating WP's in flight - it is sort of what I thought you would be doing!

I guess the main obstacle would be to find out how you would communicate with the Picoipilot this way without a PC but maybe U-Nav would be able to give you these details?

I used servo boosters from ServoCity and Cermark which both work fine BUT please note that the FMA ones did not work with FS-8/Picopilot. They don't seem to level shift from 3.3v to 5v!

And finally thanks for the tip regarding my twisted Viking wing - I managed to striten it finally and it took away all the trim needed!

/Icebear

[2bent]

Luckily the FMA servo boosters work great, enable and other commands function normally!

We've done a very successful test with a NAV only flight, but are now having trouble getting the pitch (ALT hold) control to function. The first time we enabled it, the plane snapped straight up, then straight down; and after decreasing gain, it just descends in a gradual 'stairstep' pattern instead of holding altitude.

We suspect that the elevator control needs to be reversed, but the Picopilot manual isn't too clear on how to do this, it mentions something about replacing jumpers, but so far hasn't solved the problem...

Glad you got the Viking flying well!

-Ben

BTW- we are using a Dell Laptop as a ground station, so interfacing with a wireless modem shouldn't be a problem on the ground side...

[kd7ost]

On reversing the direction of the elevator all you do is to remove or install the jumper. Installed its one direction, removed it reverses. Page 12 in the UNAV Pico Pilot manual shows you where the alt servo reverse jumper is and how it should deflect the elevator upwards in the power on test for about 1 second. If it deflects downward, the jumper needs to be removed.

Whenever you enable the unit, the stick (servo) position is read in the PIC chip to set center. What this means is that you can't be holding any elevator position up or down when you enable the unit. Have the plane trimmed for straight and level flight, hands off the sticks, then enable.

Dan

[icebear]

Ben,

Good to hear the FMA boosters worked - they must have updated them since I got mine!

Dan has already answered the AltHold question - it certainly sounds like you should check the servo direction like Dan suggested.

/Bjorn

[2bent]

Our boards have 3 jumpers each, a red one and two black ones... but our manual simply mentions 'removing THE jumper to change servo direction'. Tried removing the red one, and the elevator does deflect up in the power on test, but the plane still does an erratic 'stairstep' descent when the autopilot is enabled in flight.

Also, with autopilot enabled the elevator does not respond when the control stick is moved more than 50%. (elevator works fine when autopilot disabled, and the rudder override works fine)

I'm sending an inquiry to U-nav to see what they think...

And, the radio modems arrived (decided to buy straight from maxstream), so we'll be running some ground and airborne tests shortly :)

Thanks guys for the help!

-Ben

[kd7ost]

The two black jumpers are for programming the unit at UNAV for software upgrades. The red one on the left side is for the reversing.

Since you're using a SS for testing and it has that huge barn door elevator, I suspect you have way too much throw in it. You have to tame down the aircraft controls quite a bit to get these things to work. I use pretty small elevators and rudders on my UAV aircraft and still minimize throw. Imagine if you were steering a car. An inch either way with the steering wheel causes full wheel travel. That's just too much control to be smooth.

You have to understand the pilot is highly anticipatory and fluid. The Guidance packages are reactive. They feed in control once they detect they are coming off that line. In order to smooth that out you need a plane that is very stable and is a gentle flyer. Get rid of as much mechanical elevator travel that you can through the linkages and then try it again. The SS isn't easy to work with. It has motor thrust issues etc that all have to be dialed in before the plane will fly well.

Did you buy this used from somewhere? The Pico Pilot comes with a CD with the manual, and integration guide, waypoint editor etc. Those are helpful if they are read.

I can't say why the elevator doesn't respond when over 50 percent of stick is used. I will state that yours might be but it depends on what you're looking for. If you put in more than 50 percent stick, you will get a very slight deflection in that direction. It won't give you your full deflection by any means. It's enough to gradually change your altitude in a smooth manner. Also, sitting static on the bench doesn't net you the same results as in flight. The sensors need to be sensing changes.

Is your unit mounted slightly nose down in your aircraft?

Dan

[d_wheel]

Also, with autopilot enabled the elevator does not respond when the control stick is moved more than 50%.

-Ben

It sounds as if you might have your servo travel adjustment set to low. Try to keep travel adjustments as high as possible, around 90 - 100%. If it is set too low, it will not be enough to trigger the auto climb/descent feature of the autopilot. Then, as Dan suggested, set up your mechanical adjustments to get the amount of elevator travel you need.

Later;

D.W.

[2bent]

Reducing mechanical elevator throw helped a lot-- now we can hold altitude accurately. The flat section of the track below (Google Earth) was when the autopilot was enabled.

We've gotten our 9XStream radio modems functioning, and can recieve GPS data, as well as upload waypoint lists remotely, but are still working out a switch to do both on the same flight.

Dan- We bought straight from U-nav, and the unit is mounted slightly nose down, on a Hitec 'flight-saver' foam pad.

Thanks everyone for the help! This project is certainly fulfilling its intent as a learning experience!

http://i2.photobucket.com/albums/y2/2bent/altholdhoriz.jpg

[Munin]

Hello

The backdraw on your in flight waypoint reprogramming is that the 32'nd waypoint is autoset to the planes current position. Hence loosing return to home and non stability while reprogramming. But then again a semi-workaround is to everytime set the 31'st waypoint it the operators position. The time window while reprogramming must be as short as possible or else you might get a spin resulting in a *splat*

/Munin