Autonomous model aircraft using either BS2 or javelin stamp

Comments

• Bean Posts: 8,129

  2005-03-24 14:36 edited 2005-03-24 14:36

  I wouldn't discount the SX processor though.
  More RAM and tons of speed.
  And for the price of them you could have several in the plane.
  And with SX/B the programming would be too bad.
  
  Maybe even some SX's to handle the accelerometers and servos.
  Then the SX could "talk" to a BS2 that would handle the flight-path.
  Bean.
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  "SX-Video Display Module" Available Now.
  
  www.sxvm.com
  
  "A problem well defined, is a problem·half solved."
  ·
• allanlane5 Posts: 3,815

  2005-03-24 14:54 edited 2005-03-24 14:54

  You probably won't be able to control an aircraft only with accellerometers. Depending on what you want it to do, you may need GPS to keep track of position, wind-speed, and ground-speed.
  
  Stability requires keeping track of attitude to the ground. It's perfectly possible for a human pilot to put the plane into a spiral down, where it feels exactly like he is sitting flat and level. That's why instrument flying is so important, and so difficult. Airplanes do this with gyroscopes, I believe.
  
  The SX is probably a better bet. If you run it at a low speed (4 Mhz or so) it doesn't use a lot of power, and it has lots of RAM. In all fairness, most model planes are flown via 6-channel RF controllers from the ground, which is MUCH easier. The operator on the ground can then visually see the attitude of the aircraft, its current position, estimate its speed, etc.
  
  Even the famous UAV's being used in Afghanistan and Iraq (Un-manned Aerial Vehicles) are not autonomous. They are controlled by a guy sitting at a workstation with a video feed from the UAV.
  
  And the Javelin would definitely be a better way to go. MUCH better debugger, more powerful language, much more memory.
• glaso Posts: 38

  2005-03-24 15:21 edited 2005-03-24 15:21

  here you can find a lot of information about autonomous flying http://autopilot.sourceforge.net/
• Paul Baker Posts: 6,351

  2005-03-24 15:56 edited 2005-03-24 15:56

  I agree with allan on the GPS issue, it is the most effective means to aquire your 3D coordinates. I would also have a calibration step to set what the altitude offset is (you don't want the plane to think sea level is ground when your standing 100' above it). www.sparkfun.com has GPS units small enough to incorporate on a plane.
  
  I would include a 2D tilt sensor to aquire your pitch and roll, yaw could be inferred through your GPS cooridinates over time.
  
  Are you planning on incorporating an ultrasonic ranger finder, or any other object detection system?
  
  I would most definately incorporate a traditional RC system in it, especially at the beginning. It should be set up that whenever the receiver is getting a signal from the transmitter, the autonomous functions are overriden. Then you would switch off your transmitter to enable the autonomous functions. This is nessesary, because if you do this correctly, you will have several hundreds of dollars worth of equipment on the plane, and it would truely suck if a glitch in your programming would cause it to nose dive into the ground on your first trial run.
  
  Post Edited (Paul Baker) : 3/24/2005 4:54:01 PM GMT
• allanlane5 Posts: 3,815

  2005-03-24 16:24 edited 2005-03-24 16:24

  I once heard a model train enthusiast say he got into trains from model airplanes because a new model airplane was merely one that hadn't crashed yet.
  
  That's always bothered me about planes -- if something goes wrong, there's no place to pull over.
• voodoofish Posts: 67

  2005-03-25 02:43 edited 2005-03-25 02:43

  I would wonder about the specs of the aircraft you want to use considering the electronics are going to weigh the plane down quite a bit...
  Also do you have expirence with flying model aircraft? I imagine you would, but if not, I'd take a look into it....I've gone through 2 so far just learning..... [noparse]:)[/noparse] but it's been fun....I can't imagine how it would work with an robitic version especially when you hit a gust of wind and were thrown off course....or had a strong head wind....I've had a tough time with that with me flying!!!!
  
  Anyway just a thought on the setup... a 3 channel setup would work well for starting, throttle, rudder and elevator....simple is good....grab the bluetooth module and use one of those bluetooth gps units. I have one for my ppc , they are small and light....
  
  I'm looking to add the bt gps to my land based unit....when I build it though I think the coding is going to be tough [noparse]:)[/noparse]
  
  anyway, good luck with it....
  Ciao,
  mark
• Paul Baker Posts: 6,351

  2005-03-25 16:07 edited 2005-03-25 16:07

  Honestly, choosing the right parts doesn't add much weight, the Lassen IQ GPS is 6.5 grams, a tilt sensor would be a couple grams, Basic stamp or javalin a few grams, everything else (except object avoidance sensor) would already be on an airplane. (youd need a 3.3V regulator for the GPS but thats under a gram)
• voodoofish Posts: 67

  2005-03-25 18:57 edited 2005-03-25 18:57

  most of the bt gps units that I've seen have a battery and last for up to 8 hours or more.....long after the aircraft batter had died...so no worries on those [noparse]:)[/noparse] But with the extra batteries for the electronics and the shift in weight, I could see it being a problem for such a small aircraft. Too much weight in the rear of the plane and you've got a tail dragger that is hard to control....and you'll be wasting a lot of power trying to avoid stalling....
  
  problems I see running into:
  weight distribution on a small aircraft
  ESC hookup
  BEC if running from one battery though I'd think this would be a dangerous route to take.
  needing to know the state of the batter for the aircraft motor to know when you are running low, and have enough power to land.
• cira Posts: 2

  2005-03-25 19:23 edited 2005-03-25 19:23

  Thank you all for the positive feed back
  This project will have to be staged into at least five prototypes and the initial * work area version* would use three accelerometers and a compass app mod... this would basically be a feasibly study and it would even be probably on a glider chassis.
  
  There was a good comment that accelerometers would be faulty under several real world conditions.
  
  The gps unit would indeed be a needed addition and I think that it would defiantly be on at least the second build.
  
  Range finding /object detection systems: They would be nice for closer to end level prototypes but I would wonder if the javelin stamp can handle all the features I would be asking of it. That is nothing to say about the amount of code this will need.
  
  I think that an over ride remote system would indeed be needed, does anyone have suggestions to manage this? I mean how do I tell the plane to ignore the Auto components?
  
  Also what is this SX processor I keep hearing about?
  
  I think at least one stage of the prototypes will be merely attempting to find level flight and until I can manage to make level flight and landing possible battery life isn’t as much of a concern as it would have crashed long before batteries run out if I’m not careful…
  if you all would be so kind as to keep giving suggestions id appreacate it
  once again thanks for the interest and support
• Paul Baker Posts: 6,351

  2005-03-25 19:33 edited 2005-03-25 19:33

  The winners of the SX design contest were just announced, and guess what the 3rd place project was? Thats right, an airplane autopilot, heres a link to that project: http://www.parallax.com/sx/contest/contest_airplane_autopilot.asp
• voodoofish Posts: 67

  2005-03-26 00:30 edited 2005-03-26 00:30

  Very nice, I'm going to have to look at the report...
  
  Also I thought this might be of interest...
  
  http://www.rconline.ca/products/mi-soft/rcap/
  
  it was designed for the pic. controller, but I'm sure it could be translated somehow to use the basic or java code instead....
  he also could connect to a gps.....
  
  Enjoy
  Mark
• voodoofish Posts: 67

  2005-03-26 01:56 edited 2005-03-26 01:56

  So where can I find one of those ISM modules???? He doesn't mention a site for that.....and it sounded like it would be useful for getting all sorts of sensor data....
• phuzzy3d Posts: 3

  2005-03-28 06:23 edited 2005-03-28 06:23

  members.shaw.ca/sonde/index.htm
  
  
  you may find his site very useful.
• Paul Baker Posts: 6,351

  2005-03-28 15:40 edited 2005-03-28 15:40

  Here's a link to an ISM module: http://www.radi.com/modular43.htm·if you look at his board you'll see its not the same one (his is an off board antenna), but using the information in the link and google *nudge* you can find other vendors, and perhaps the one that made his.
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
• voodoofish Posts: 67

  2005-03-28 21:53 edited 2005-03-28 21:53

  I tried google and yahoo...and found a lot of various parts, but nothing similar and with a base station for recieving the info...
  maybe I'm looking up the wrong things, but I've tried at least a dozen different searches. When I've found something, I can't find a distributer....
• Paul Baker Posts: 6,351

  2005-03-28 22:19 edited 2005-03-28 22:19

  frickisnickiswicka di, why why why why does this thread always throw an error when I try to post to it? I am livid about this, I shouldn't have to quick post, then go back and edit it di!
  
  sales@radi.com
  
  http://www.lyonscorp.com/components.html·-> http://www.lyonscorp.com/distributor_list.html
  
  you'll need to put effort in aquiring one, no online sales outlet, you can always inquire with the guy who used it.
  
  ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
  
  
  Post Edited (Paul Baker) : 3/28/2005 10:26:10 PM GMT