Post by thread:interesting project

>From another list, some might find it fun to look azt. Derward ----- Original Message ----- From: "Tyler Harpster" <spam_OUTtyler881TakeThisOuTcomcast.net> To: <.....microwaveKILLspam@spam@lists.valinet.com> Sent: Monday, April 27, 2009 10:23 PM Subject: [Mw] microwave pointing device ideas {Quote hidden}>I was approached by an astronomer for a device to point north that does not > rely on a magnetic compass. He needs to be able to accuratly point north > without also seeing the stars. (cloud cover while setting up in the hopes > of a break when you need it) > > The microwave crowd also needs an accurate azimuth direction indicator, > right? > > Anyway.... this is the best I could come up with. I'm waaaayyyy too busy > to > build something like this, so if someone wants to run with it and turn it > into a product... have at it. (if it's feasable) > > Here's the idea: > > Product would be about the size of a shoe box? > > Cheap GPS receiver pumps out NEMA data every second. (faster would be > better) > That data is sent to a microprocessor. > WHILE you are driving on a relatively nice road with few trees by the side > of the road, (for good GPS coverage) > the microprocessor continually takes GPS points and continually > averages > your direction of travel. (some algorithm work here as averaging GPS data > points is an art all to itself) > Now that the microprocessor has a continual average of direction of > travel... > A Gyro from a RC helicopter is used to continually store the direction of > travel and report any changes made to it back to the microprocessor. > (more > algorithm work-determining when to stop accepting GPS inputs) No, I > don't > know how accurate gyros from RC helicopters are. Sorry. It would have to > be pretty accurate to work for this application. > > So after a few miles at highway speeds, the box would "settle out" and > converge on your exact direction and each input you make to it. As you > slow > down, the box would stop accepting GPS inputs because they would be too > close together. When you come to a stop, the gyro would continue to hold > the true direction data and you'd take the box over to your tripod and > mount > it. (unclear of what the mounts look like) The box would have to have a > true side on it like a level or square and indications for which direction > it is currently pointing. > > Ideas? Thoughts? > > Should I go back to the drawing board? > > Tyler > > _______________________________________________ > Microwave mailing list > microwaveKILLspamlists.valinet.com > www.valinet.com/mailman/listinfo/microwave >

Lots of projects out there that have information on how to do this, and the expected accuracy. Look up Kalman filters and IMU (inertial measurement units). Perhaps start with this: http://www.sparkfun.com/commerce/product_info.php?products_id=9184 They have an IMU section: http://www.sparkfun.com/commerce/categories.php?c=85 Might also look up ardupilot, and the many forums for electronics, navigation, RC vehicles, etc. -Adam On Tue, Apr 28, 2009 at 9:41 AM, Derward Myrick <.....wdmyrickKILLspam.....earthlink.net> wrote: > >From another list, some might find it fun to look azt. > > Derward > > > > {Original Message removed}

How well this will work depends a lot on the application. There are no gyros out there which will allow you to "set" the orientation and then have it be accurate for an indefinite amount of time thereafter. You will need some way to re-set the gyro periodically. A good GPS receiver should be able to get heading info from even fairly slight motion - you don't need to drive it down a road. Another possibility would be to use radio direction finding to find the bearing to several known radio transmitters (such as FM radio stations - could even use time-difference of arrival for this) or aviation navigation beacons. Would this device be used to determine bearing angle in one location or would it have to be able to do so anywhere on earth? Sean On Tue, Apr 28, 2009 at 11:42 AM, M. Adam Davis <EraseMEstienmanspam_OUTTakeThisOuTgmail.com> wrote: {Quote hidden}> Lots of projects out there that have information on how to do this, > and the expected accuracy. Look up Kalman filters and IMU (inertial > measurement units). > > Perhaps start with this: > http://www.sparkfun.com/commerce/product_info.php?products_id=9184 > > They have an IMU section: > http://www.sparkfun.com/commerce/categories.php?c=85 > > Might also look up ardupilot, and the many forums for electronics, > navigation, RC vehicles, etc. > > -Adam > > > On Tue, Apr 28, 2009 at 9:41 AM, Derward Myrick <wdmyrickspam_OUTearthlink.net> wrote: >> >From another list, some might find it fun to look azt. >> >> Derward >> >> >> >> {Original Message removed}

Typically a moving type GPS (not a land surveyors or other using DGPS local station, or averaging over a period of time) will be +/- 30 feet or so. Lets take an extreme, 60 foot error between readings, along with 1000 foot of travel, then arc tangent of 60/1000 = 3.4 degrees. But the unit will be taking many readings in a minute. WAAS (or the Russian equivalent) will be a little better. Some units may average the track direction. I have used the unit to find cemetery grave markers, after previously setting a waypoint to the marker with good success, in particular when after vising several times, each time adjusting the waypoint location by averaging the old and new locations. I use the map then to determine what direction to walk. Sorry to say I haven't paid attention to the track direction indicated, or whether WAAS was active. I have a Lowrance Expedition C hand held. For short distances of travel, say maneuvering in a small parking lot or small street intersection, trying to figure out the direction of travel, or where you want to go is not dependable. The unit does have a built in compass which works fine when a passenger of the foot (walking), but inside a vehicle, is useless, with errors of 90 degrees or more, even after compensating in the vehicle. Sean Breheny wrote: > A good GPS > receiver should be able to get heading info from even fairly slight > motion - you don't need to drive it down a road. >

Hmmm, I was just going by experience with a Garmin consumer grade hiking GPS about 5 years ago. You could literally hold it in your hand, walk a few steps, and it would give you a reasonably accurate heading (say, within 20 deg, although I didn't rigorously test it) A GPS fix has several types of error (this is a rather generic analysis): 1) zero-mean noise on the time-scale of seconds 2) zero-mean noise on the time-scale of minutes/hours 3) a constant (over an hour, say) mean error which depends upon your position, which satellites you see, etc. A single position fix has all three sources of error. Averaging over seconds reduces #1, averaging over hours reduces #2, DGPS partially allows you to eliminate #3. A velocity estimate isn't affected by #3 at all. Some sensible averaging and assumptions about how you can move can be built into the GPS's filtering algorithm so that #1 and #2 are reduced. This leads me to believe that a velocity estimate has the potential to be quite accurate. Sean On Tue, Apr 28, 2009 at 7:37 PM, Carl Denk <@spam@cdenkKILLspamwindstream.net> wrote: {Quote hidden}> Typically a moving type GPS (not a land surveyors or other using DGPS > local station, or averaging over a period of time) will be +/- 30 feet > or so. Lets take an extreme, 60 foot error between readings, along with > 1000 foot of travel, then arc tangent of 60/1000 = 3.4 degrees. But the > unit will be taking many readings in a minute. WAAS (or the Russian > equivalent) will be a little better. Some units may average the track > direction. I have used the unit to find cemetery grave markers, after > previously setting a waypoint to the marker with good success, in > particular when after vising several times, each time adjusting the > waypoint location by averaging the old and new locations. I use the map > then to determine what direction to walk. Sorry to say I haven't paid > attention to the track direction indicated, or whether WAAS was active. > > I have a Lowrance Expedition C hand held. For short distances of travel, > say maneuvering in a small parking lot or small street intersection, > trying to figure out the direction of travel, or where you want to go is > not dependable. The unit does have a built in compass which works fine > when a passenger of the foot (walking), but inside a vehicle, is > useless, with errors of 90 degrees or more, even after compensating in > the vehicle. > > Sean Breheny wrote: >> A good GPS >> receiver should be able to get heading info from even fairly slight >> motion - you don't need to drive it down a road. >> >