Jump to content


Photo

Skyhunter UAV Build Log


  • Please log in to reply
56 replies to this topic

#21 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 10 February 2013 - 03:21 PM

Brief update: Build is now complete, and was test flown on Friday. Autopilot is working flawlessly, it's been a great success so far. Will update with pics in the next couple of days.
The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#22 warmachine

warmachine

    Overlord

  • Atomican
  • 325 posts

Posted 10 February 2013 - 04:12 PM

what you study to be able to do this?

#23 Guest_MrInsaneBuff_*

Guest_MrInsaneBuff_*
  • Guests

Posted 10 February 2013 - 09:24 PM

what you study to be able to do this?


Awesome. He studied Awesome minoring in Freaking Insane.

#24 warmachine

warmachine

    Overlord

  • Atomican
  • 325 posts

Posted 11 February 2013 - 05:21 PM

yea i think this technology is of great importance. Since the planet will soon be finished there has to be robots to scan the damage area

#25 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 11 February 2013 - 11:45 PM

what you study to be able to do this?


I studied mechatronic engineering (robotics). Although it makes things a lot easier for me, just about anyone can set up a drone these days, which is both awesome and a bit terrifying. Check out www.diydrones.com for more info if you're interested.

Another update: Back in civilisation (i.e. a place with internet). I'm rendering the DEMs, which is taking a bit longer than expected. I'm processing 221, 12 megapixel images which all overlap by roughly 70%. It's taken 5 hours on an old 2.8Ghz Core i7 machine to create the initial point cloud, and I've had to break it up into smaller segments due to memory limitations (I only have 8GB to work with). Smaller sections rendered on my laptop were looking good, with a surprising amount of detail captured using low resolution settings.

The build turned into a lesson on self indulgence and perfectionism. The final product was horrifically over engineered, and it took a weeks worth of late nights and early morning to get it prepped in time to fly. I need to sort out the photos into something that tells a coherent story, but I was quite pleased with the results. I've got a couple of rest days before the next mission, so will add to this thread as I get through them.
The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#26 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 18 February 2013 - 01:13 PM

I'm going to break the build log up into various components, rather than trying to capture an accurate timeline in the correct order. It was a hectic week of building, and there was room for improvement in my documentation skills. I'm back in China now, and I'm happy to report that the project was an overwhelming success.

Camera Mount

Even with electric flight, there are still significant sources of vibration that can degrade aerial photo quality. The propeller can be statically balanced, but I don't have the sophisticated equipment necessary to dynamically balance it. Additionally, I'm using a hobby grade electric outrunner motor, which in an ideal situation, would also needed to be dynamically balanced. To overcome these limitations, I built a specialised vibration mounting camera mount.

Posted Image

It's based on these custom rubber grommets, which are designed to dampen vibration in the 5-10kHz band. The baseplate was made of light ply.

Posted Image
Posted Image
Posted Image

4 of them together is enough to support the weight of the camera. A second light ply plate is fixed to the top to provide a platform for the camera.

Posted Image

2 layers of foam provide final isolation. The first is firm, the second much softer. Pegs were added to the side to allow the camera to held to the top plate using rubber bands.

Posted Image

The next task was to roughly fit everything together to determine where the batteries were needed to get the correct centre of gravity required for stable flight. I considered mounting the camera on an external pylon for some time to allow for easy access, but found that the batteries were going to be sitting far enough back that I could mount the camera right under the access bay.

Posted Image

After measuring everything carefully, I cut a hole in the bottom of the fuselage. To get a nice cone shape, I made a jig that fitted to a cordless drill with sandpaper stuck to balsa wood with the correct profile. Given how difficult EPO foam can be to work with, I was pretty happy with the results.

Posted Image
Posted Image

The inside was lined with a light ply plate. A 55mm UV lens protector was then mounted to allow for an undistorted window that woud protect the camera during belly landings.

I used a vibrometer app on my smartphone to measure the effectiveness of the mount. First test had the phone resting on the bottom of the fuselage, and the second had it rubber-banded to the mount. High frequency vibrations were just about eliminated, and low frequency vibrations were reduced by ~90%.
The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#27 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 18 February 2013 - 01:37 PM

Airframe

Posted Image

I modified the tail of the Skyhunter to include rudders. Later on, I found out the reasons the manufacturer ditched them, but it did make flying in cross winds better. I also added some carbon fibre reinforcement to account for the additional weight and twisting forces the addition of the rudders had.

Posted Image

The GPS was mounted just behind the canopy, with a balsa plate to streamline and secure it.

Posted Image
Posted Image

Aileron servos mounted in the wing, and cables to power the elevator and rudder servos routed through to the booms.

Posted Image

The fuselage halves were joined, and the inside lined with velcro. The allowed flexible placement of batteries and other hardware to account for different flight configurations. You can see the ESC (motor controller) and BEC (power supply) mounted at the rear, and the current and voltage sensor for monitoring the batteries mounted on the left.

Posted Image

The data radio was mounted under the wing. This was less than ideal, as it made access difficult, but it was a compromise that needed to be made. There are 5 radio transmitters/receivers onboard - GPS receiver for navigation, GPC receiver on the Camera, RC receiver for manual control, data radio transceiver, and finally the video downlink transmitter. Each of these is either sensitive to or emits RFI, and needs to be spaced appropriately and away from other sources of RFI, such as the ESC, BEC and Motor.

Posted Image

The APM and RC receiver was mounted on a removable light ply tray. Some anti-vibration foam was used to reduce high frequency noise.

Posted Image
Posted Image
Posted Image

Finished! Vinyl was added to make it more visible. There's quite a bit that I didn't document, but this was most of the interesting stuff. Things like mounting the pitot tube in the nose to measure airspeed, and endless hours of small and frustrating bits.
The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#28 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 18 February 2013 - 02:06 PM

Motor Testing

Posted Image

Before flying, I needed to calibrate the current and voltage sensor, and choose an appropriate propeller size to match the engine. All of the combinations were simulated using eCalc, so the main purpose of the test was just to confirm the theory. Everything went well, and the engine was producing about 2.6kg of static thrust using a 12" x 6" propeller, and drawing around 42 Amps.

Edited by tastywheat, 18 February 2013 - 02:06 PM.

The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#29 AccessDenied

AccessDenied

    Apprentice

  • Hero
  • 11,839 posts

Posted 18 February 2013 - 05:37 PM

I go away from Atomic for a few days and miss this.

Been ages since I've seen you TW. And I don't think I'm alone in having a man-crush on you! This is awesome.

Following with great interest.

Brought to mind this:
http://www.enchanted...e/archives/1389

*snip*
But a new method of fighting poachers has now been introduced in the 90,000 acre Ol Pajeta Wildlife Conservancy. The innovative approach, which will be the first of its kind in Kenya, makes use of an unmanned plane that will be equipped with a camera to track white rhinos and other threatened animals in the conservancy using Radio Frequency Identification (RFID) tags.
*snip*

And for those interesting in donating money or wanting to make a difference somehow:
http://www.kws.org/

Kenya Wildlife Service.

This is the government group that trains Rangers to protect poachers. This is not WWF protecting Pandas stuff. This is making sure that the KWS has guns, ammo, bullet-proof vests etc so they can take the battle to the poachers. There is even an option to donate money to widows of the rangers.

Wildlife poaching is serious business. Chaining yourself to a tree will just get you shot there.

You have a noble task with your plans TW.

AD
I can make my own woggle! Bet you can't woggle yours.

#30 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 19 February 2013 - 02:25 PM

Brought to mind this:
http://www.enchanted...e/archives/1389

*snip*
But a new method of fighting poachers has now been introduced in the 90,000 acre Ol Pajeta Wildlife Conservancy. The innovative approach, which will be the first of its kind in Kenya, makes use of an unmanned plane that will be equipped with a camera to track white rhinos and other threatened animals in the conservancy using Radio Frequency Identification (RFID) tags.
*snip*


I'm pretty familiar with the security guys at Ol Pejeta, they're all awesome people who have been given an incredible responsibility. There's a species of Rhino known as the Northern White Rhino that is beyond critically endangered. There are only 6 individuals left alive that we know of, and Ol Pejeta looks after 4 of them. They were all born in the Czech republic into captivity, and I had the privilege of being at Ol Pejeta in 2009 when they were brought over as a last resort to get them to breed. I got to see them stumble around after the travel sedatives had worn off, and take their first bites of African grass. It was one of the experiences that lead me towards my current path.

They raised the money for the UAV via IndiGogo, which is similar to what I'm hoping to do. The campaign is now closed, but there's still some interesting information on their page: http://www.indiegogo.com/olpejeta
The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#31 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 19 February 2013 - 03:25 PM

Stage 3: Flight Testing

With the final pre-flight checks done, I took the plane to a local aero-modelling club to perform the test flights. It's a nice strip near Mareeba, with flight clearance to 1,000ft and lots of clear space. Weather was less than ideal, with winds gusting to 25km/h. With the range checks done, the expensive camera gear stripped out, and the GPS locked onto 12 sats, my Dad gave it a good hand launch and it was airborne. With a thrust to weight ratio of about 0.75, the initial climb was authoritative and smooth. There was some very minor trim needed in the elevator and ailerons to get in to fly straight, but after that, even in the gusty conditions, it was flying with great stability. Stall behaviour was decent, with a slight tendency to drop the left wing if pushed too hard. Landing required a fair bit of space because of the relatively high wing loading, but was otherwise uneventful.

Posted Image

Groundstation setup. Tools everywhere, laptop shielded from the sun using a Sony sun shade I'd found on Taobao, the Chinese equivalent of eBay.

Posted Image

Taken just after it had been hand launched.

Posted Image
Posted Image

Colour scheme showing up fairly well.

Posted Image

Landing.

After a brief first flight, the trims were transfered to the Autopilot, and a basic set of waypoints were set up. Control loops were set up for loose control, and a 3-position switch on the transmitter was used to switch to change from manual, stabilised and auto flight modes. The stabilised flight mode was the first to be tested. It aims to hold the aircraft straight and level, and will co-ordinate turns with a simple aileron input. I flew the plane manually to about 100m altitude to give a bit of room for mishaps, and then flicked the switch. The plane went from being knocked around by the wind under manual control, to flying as if it were on rails. There was a slight wobble in the turns, which suggested my kp (proportional constant of the PID controller) was a bit too aggressive, but that was easily adjusted mid-air via the GCS.

Then next thing to do was to try out Auto. 4 waypoints in the basic shape of a circuit were uploaded while the UAV was still in flight, and the Auto switch was flicked. A computerised voice came from the GCS "Mode changed to Auto, heading to waypoint 1" and just like that, the plane was flying itself.

Posted Image

It was quite at eery feeling watching the plane fly itself, and it took a bit of time before I was confident that it wasn't going to fly away on a whim. The blue squiggles are where I was flying it manually. The wind was blowing it off course quite a bit between waypoints 3 and 4, but I was able to improve the tracking by fiddling with the navigation routines and altering the tuning of the cascading PID loops.

With a successful test flight done and dusted, it was time to have some fun with FPV. I mounted a basic camera to al old model I had built out of depron that was designed to fit in a briefcase. I had taken it with me to Kenya in 2009, and had fun chasing Zebra around with some basic cameras set up to film what was below.

Posted Image

Fixing the video transmitter and camera to the fuselage.

Posted Image
Posted Image

Attempting to land via the video feed to the goggles.

It was good fun, and I confirmed that the range of the video downlink was far beyond what was comfortable to track visually. The FPV makes it just like flying a real aircraft, which is surprisingly challenging when you are used to flying RC. Judging airspeed, co-ordinating turns and navigating are much for difficult when you don't have a 3rd person frame of reference.
The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#32 martyr

martyr

    Master

  • Atomican
  • 1,094 posts
  • Location:Australia

Posted 19 February 2013 - 04:06 PM

Man, all of this is insanely cool! Congrats!
Q6600@3.00Ghz/Prolimatech Megahalems w/ Coolermaster 2000RPM 19Dba FAN/Corsair 4x2GB 800Mhz DDR2/Gigabyte GA-EP45-UD3P/Inno3D GTX580 1.5Gb/OCZ Agility 3 240Gb/Corsair TX-750/Antec 900

#33 warmachine

warmachine

    Overlord

  • Atomican
  • 325 posts

Posted 19 February 2013 - 07:09 PM

yeah definately the best thread I have ever seen in my history of atomic. Maybe because I didnīt know I was interested in this

#34 chrisg

chrisg

    Learner

  • Super Hero
  • 30,809 posts
  • Location:Perth

Posted 19 February 2013 - 07:57 PM

:) Cool is right, very, very cool ;) Just an idle thought, not trying to rain on your parade, but in pursuit of poachers how vulnerable is it going to be to say a shotgun blast? I'm assuming that might relate somewhat to how quiet it is and to having a low viz underside? Cheers
"Specialisation is for Insects" RAH

#35 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 19 February 2013 - 11:53 PM

Just an idle thought, not trying to rain on your parade, but in pursuit of poachers how vulnerable is it going to be to say a shotgun blast?

I'm assuming that might relate somewhat to how quiet it is and to having a low viz underside?


It would be as vulnerable to a shotgun blast as a passing bird, so it's definitely something to consider. The poachers tend to have assault or very large calibre rifles rather than shotguns, which would make the task a bit harder. Some are simple opportunists, though many are former KWS rangers or ex-military. They're organised, and generally well trained, so I'm sure they would eventually find a way to take them out if they wanted to.

There's more to consider than just the airframe though. The first alert to poachers in the area are usually gunshots. By then, the animal is usually already lost, and the poachers run off with their bounty by the time rangers can respond. By firing at the UAV, they risk giving away their position. If they manage to shoot it down, they're essentially telling the rangers where they are to a proximity of a couple of hundred meters. Assuming it's carrying a thermal camera (the most expensive bit of equipment), you've paid ~$10k to locate some active and armed poachers. Not such a bad deal.

Secondly, Africans are very susceptible to mythology. Many don't have a good understanding of technology, and tend to view it in organic ways. They would view a plane with no pilot as something like a phoenix, and it would likely scare the shit out of them. A strange angular animal that can see in the dark, never eats or sleeps, and spends it's time hunting them down. I plan to utilise this aspect as much as possible as a deterrence.

Edited by tastywheat, 19 February 2013 - 11:54 PM.

The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#36 chrisg

chrisg

    Learner

  • Super Hero
  • 30,809 posts
  • Location:Perth

Posted 20 February 2013 - 08:31 AM

Very interesting TW. Thanks for the response. BTW, don't know if you are an IEEE member but they have a lot of interest in UAVs, you'd have the makings of an article in Spectrum magazine right here. Best of luck. Cheers
"Specialisation is for Insects" RAH

#37 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 27 February 2013 - 06:55 PM

Adventure stories update: I returned to work to find out I had been promoted to chemistry teacher. Everything in China is done at the last minute. This caught me off guard, as I don't have a background in Chemistry, and I'm now expected to teach 76 year 12 students the South Australian chemistry curriculum (SACE exports their curriculum to overseas schools). In 9 months. Not one to back away from a challenge, I've been busy studying my arse off, hence the lack up updates here. I'm rendering the DEM to a .kml file, so those with Google Earth can have a detailed look around. Will upload here when it's done, the render is 34% complete with 6 hours remaining.
The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#38 Cybes

Cybes

    Apprentice

  • Atomican
  • 15,251 posts
  • Location:Where I am

Posted 27 February 2013 - 09:23 PM

Year 12 chem on 9 months notice... After careful and prolonged consideration, here is my verdict: Fuck. That. Good luck, TW - that's going to be a challenge if any of your students are switched on.

"Reality does not care what you think." - Dr Richard Feynman

"If you want to build a ship, don't drum up the men to gather wood, divide the work and give orders. Instead, teach them to yearn for the vast and endless sea." - Antoine de Saint Exupery [quoted by Niel DeGrasse Tyson, about space travel]


#39 tastywheat

tastywheat

    Primarch

  • Atomican
  • 903 posts

Posted 28 February 2013 - 11:24 PM

After careful and prolonged consideration, here is my verdict: Fuck. That.


I'm coming to the same conclusion. Things have become considerably conflicted after I started asking questions and pointing out flaws in the plan. Lesson learned: Don't call out bullshit in China, especially when the person spouting it has guanxi (literally translates to connections, but essentially means socially acceptable corruption). I've been here long enough to know better, but it was hard to resist in this case.
The woods are lovely, dark and deep. But I have promises to keep, and miles to go before I sleep -Frost

#40 Cybes

Cybes

    Apprentice

  • Atomican
  • 15,251 posts
  • Location:Where I am

Posted 01 March 2013 - 12:26 AM

Things have become considerably conflicted after I started asking questions and pointing out flaws in the plan.

The gentle art of bullshido, aka verbal judo: making it obvious to all and sundry that someone is talking crap by letting them verbally hang themselves in public?

"Reality does not care what you think." - Dr Richard Feynman

"If you want to build a ship, don't drum up the men to gather wood, divide the work and give orders. Instead, teach them to yearn for the vast and endless sea." - Antoine de Saint Exupery [quoted by Niel DeGrasse Tyson, about space travel]





0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users