Comments

• 

  Stephen Gloor March 10, 2013 at 4:33am

  @James - I didn't think that you were negative.  It is a huge commitment and I was making a way for all of us to participate with to a varying degree.  Your imaging experience and coding will be very valuable and I wanted to make sure we can take advantage of them.

  I am in Singapore next week installing network gear on a new rig.  I work offshore a fair bit but I do not do four week hitches.  I will have to work around my offshore commitments as best I can.

  I am back next weekend so maybe a Sunday is doable or one night week the week after next.  The rules should be out then.  We really cannot agree on anything until the rules are out and we make sure they are much the same as last year.

  In the meantime has everyone read Tridge's write up of last year?

  http://diydrones.com/profiles/blogs/canberrauav-outback-challenge-2...
• James masterman March 10, 2013 at 4:03am

  @Stephen, I didn't mean to sound so negative in my last post. What I meant was that I am willing to commit to the OBC and will drop my other pet projects to do it. I am happy to work on coding and anything else that is needed.

  We really need to get together to work out the basics like airframe, autopilot (I assume APM will be what we use?) and other things so we can get started.

  Are you free anytime this week or next weekend? That goes for anyone else who is interested. We could meet at a pub one evening if that makes it easier!
• 

  Stephen Gloor March 10, 2013 at 1:55am

  @James - One thing we do need is a failsafe watchdog processor to implement the recovery rules.  If the plane flies out of the safe area it is basically required to dive into the ground.  I would prefer a parachute recovery.

  If you can do it we need to program an Arduino to do geofencing.  The APM does it, if we use it, however I would prefer to set the flight controller geofence just inside the real one so we get a warning when we approaching sudden death.  The watchdog failsafe can have the real geofence and trigger the dive or parachute when the line is crossed.

  This is an example of some of the work that can be done without committing fully to the team.  We will have heaps of these things.
• 

  Stephen Gloor March 10, 2013 at 1:49am

  @James - It think we can have a couple of levels of commitment.  A person of your programming skills would be invaluable if we can machine find Joe.  My thought is to have a large computer on the ground crunching the returned frames while we also have two camera people looking as well at the real time video.  If the camera people find it first then we forgo the ten points.

  You could assist with programming without being fully committed to the whole thing.
• 

  Stephen Gloor March 10, 2013 at 1:47am

  @JB - there are two competitions.  The school's one is this year and it is only opent to high school students.  The SAR one we want to enter is in September 2014.  The rules for 2014 have not been posted yet and could differ from 2012 so we really need to wait until they are out.  They will also have the search area.

  The fact that so few teams made the search area is that it is hard.  There is an effect they call the Kingaroy Triangle that makes equipment not work.  The reality is that few people are prepared for operational use.  The sharp end of the military is often one step behind the latest because anything used in the field must be rugged and reliable not necessarily the best there is.

  My thoughts on strategy is that with modern cameras that now fly on just about every FPV plane about we can find Joe.  We may need to drop the 10 points for a machine find however if we drop accurately we still may win.

  A simple rugged airframe with a simple camera system that we have flown to death for a couple of months will be better that the latest camera on the latest and greatest airframe.  I have sketched out an airframe with twin electric motors and efficient low drag glider airfoils (SD7037/7043) that may do the job.  I will try to get this built and see how it goes.  If we decide fuel it is hard to go past the Mugin.

  I agree it is hard to fly multiple sorties in the one hour however if our plane is relatively fast we can carry and drop the dumb bomb and then come back and get the programmed smart bomb to drop.
• James masterman March 9, 2013 at 11:58pm

  I've been thinking long and hard about this one, I'd really like to enter, but the key is commitment, both mine and anyone else on the team. From the research I've done it will consume all spare time between now and the event, so if you have any other side projects, be prepared to drop them now. Because of this, it would be good to have a post comp goal in mind for the technology we develop, as others have mentioned. There may eventually be a business in it.
  
  I am willing if others are. I can bring to the table programming skills (c, c++, c# on windows including limited opencv, some ios objective-c and some arduino) also general UAV building experience with Arduplane and various foamies. I can solder ok but am not really experienced in building hardware from scratch. I've never had experience with fuel planes.
  
  I also have family in Brisbane who can lend a car and trailer and accommodation before/after the event to assist with logistics.
  
  I agree on the bottle drop comments. Is it allowable to find joe with one airframe and then send another with the bottle to the location? A fast edf could ge there quick once he was found.
  
  I suggest all interested parties meet up for coffee or even dinner one day soon to kick things off.
  
  James
• 

  JB March 9, 2013 at 11:41pm

  Stephen

  Campervans sounds like a good idea...i'd possibly drive over with our own. We also have many relatives in the Brisbane area.

  I completely agree that we need to function as a team. That includes our presentation etc.

  Am I reading it right that the delivery challenge is an additional challenge to the locations search, and the SAR search and delivery? In the delivery challenge you only have 20 minutes with setup to achieve the delivery to location? Also isn't it only open for Australian high school students...my son is in high school.

  That's effectively an "immediate" response type situation ie open the UAV ground station case, plug in the lipos and fly. There's a few minutes lost just for getting there in flight time. I'd opt for three guided bombs over dive bomb. Does the package have to land intact as well? That could be a challenge in itself.

  On propulsion: don't get me wrong i'd prefer to do everything with electric motors, but it will depend on how small we can get a reliable airframe with all the gear onboard, with enough range/endurance to find Joe. They allow up to 60 minutes so realistically we'll need to stay aloft some +45minutes unless we get lucky...the search area is 3.6km x 3.6km. That's a big area to search in with any accuracy. We can do it low and fast, slow and high etc etc. We need to find the best strategy depending on the airframe we choose.

  So if we can get it on a wing or something relatively small, electric is the only way to go. But as an example, if we need to carry 3x 500g guided bombs as well as a payload on the same airframe, I can't think of a small electric aircraft that will do it yet. Does anyone know of any +3-4kg payload electric aircraft that will not crash when you eject half of it?The "bombs" can effectively only be located on CG. Also 1.5litres of volume is not a trivial amount of space for a UAV to accommodate internally. It's nearly impossible to fly multiple sorties in the time allocated for the delivery challenge, and the "one shot" in the SAR challenge needs to be accurate to within 100m. It's obviously possible, but not with every air frame.

  However, I still think the "finding" part will be the biggest challenge. We need some good quality cameras and reliable video link to be competitive. Good cameras and equipment cost money too. What cameras can we use, how much to they weigh and cost, and do they fit on the air frame? Maybe we should start from the optics and video TX and work back to the required air frame from there? BTW does anyone know how many people are allowed to look at the video feed(s) coming from the UAV to find Joe?

  On the wiki page it says that they didn't even find and deliver to Joe last year, hence to prize winner...and only two teams of the 70 entries successfully completed the search phase at all... ooopsies. Maybe the "commited" here should do a intel collection trip to this years event? ;)
• 

  JB March 9, 2013 at 10:26pm

  Hey John that sounds good for extra long range applications. What sort of data usage do you have included with that, is it cost effective to run video? At 10km I think we should be still ok without "3G" range, but as a backup it would be worthwhile.

  System redundancy should not be underestimated in the grand scheme of things, on the day only the functional systems will make the win happen. Has anybody been thinking about search strategies?

  BTW Does anyone have the search area already so we can tell roughly how "challenging" the terrain is? 

  John is your property relatively cleared for easy automated flights? I might have a closer option just south of the Estuary I could organise, on a few hundred acres that belongs to our neighbors, actually on the Forrest Hwy. We really would need the space and lots of time to do realistic tests. We get a relatively strong sea breeze at times down there in the afternoons, but that might give us some extra skills, should the event itself be windy as well! The bonus would be that our workshop is only 10min away. There's also lake/estuary and state forest terrain within RC range with good access roads, including straight ones we could use for a chase car for testing purposes. We''ll need to start getting "first hand experience" as soon as possible I suppose. There's also a few camping spots nearby etc. for weekend stints.

  I was even thinking we could do some record range attempts out Lake Grace way, where there is also some very long straights. +50km should be possible with some planning.
• 

  Stephen Gloor March 9, 2013 at 9:43pm

  I can bring to the table logistics.  As I said before i have family with campervans in Northern NSW that will help.  We can base there before travelling to Kingaroy.  I can also do the electronics and airframes.

  The winner of the challenge is one that gets a camera to search area.  I know this sounds obvious however several highly fancied teams did not achieve this.  I think they focused on the technology rather than a robust solution that would do the job.  Though fuel has a longer range electrics are more reliable, give less trouble and have vastly less vibration. 

  Part of the challenge points are for teamwork and preparation.  We have team shirts to do, checklists etc so the team we take looks professional.  If two teams, including us, both rescue Joe then in the final count we could lose on the way we look.  Have a look at the point scoring system.

  Most of all and the most neglected area is the drop.  I am not going if we cannot reliably 100% of the time drop a water bottle within 2 meters of a target.  We can have a guided bomb or dive bomb as dive bombing is the only accurate way of delivering a dumb bomb.
• 

  John Cousins March 9, 2013 at 8:59pm

  It's a long drive... but I've just bought 20 acres 10 minutes from Mount Barker and far from any "aircraft landing zones" so feel free to use it for a test site. I have , 5.7 - 5.8 Ghz and 900 Mhz (directional helicals) that you may borrow and have some 3DR telemetry gear that i'm happy to lend. I have a VERY expensive 3G telemetry and video streaming package. I can't afford to buy it again though, so I must get it back! The 3G package uses 2 special sim cards that operate on the telstra network with a custom APN. Data doesn't travel over the internet as such (think of a private network within the telstra NextG network) and provide direct IP to IP streaming. Data can only travel between 2 fixed IP addresses that I have reserved. One in the air and one on the base station. Getting it this far was an absolute ball ache! Haven't fired it up in anger yet due to the usual lack of time but this sounds like a good application for it!