2

In my application, I need to find true north direction, for reference purpose. I am having a GPS enabled drone. I have searched over Internet and found that moving in the north direction while keeping your Longitude constant, the path followed will be the direction of true north. However, I don't know how much accurate this method is.

Can anyone please explain the best way to find the true north direction in a Drone/UAV?

Akshay
  • 43
  • 1
  • 4
  • 4
    Why not use a magnetometer and gyroscope like everyone else? – Ignacio Vazquez-Abrams Sep 12 '16 at 18:58
  • 1
    @IgnacioVazquez-Abrams So his will be the only UAV working when the pole flips? It's [not been very stable](http://maps.ngdc.noaa.gov/viewers/historical_declination/) – Spehro Pefhany Sep 12 '16 at 19:20
  • 1
    Wow, @SpehroPefhany! How have I not heard anything about this until now?!? ("wow" is a more appropriate word than my original blasphemous expression) – bitsmack Sep 12 '16 at 19:45
  • 2
    The meridians of longitude run True North/South – Peter Bennett Sep 12 '16 at 19:50
  • If your UAV operates above the clouds, and either after dark, or at enough altitude, build a star tracker. –  Sep 12 '16 at 20:03
  • 1
    @bitsmack: That's what happens when your EM field generator is a massive, churning blob of liquid within a thin, crunchy shell. – Ignacio Vazquez-Abrams Sep 12 '16 at 21:25
  • 2
    Are you looking for heading or track? In there air there is wind which can make the two rather different. – Gorloth Sep 12 '16 at 21:35
  • 2
    If the drone is constantly moving, then the GPS heading data is accurate. But if the drone is floating on a steady state position, then no GPS heading is valid. You can check the same, sitting in your car and use a GPS navigation. – Marko Buršič Sep 12 '16 at 19:21
  • A reasonable autopilot could be programmed to keep a fixed heading. As @Gorloth mentioned you need to be sure of what you want. Your plan to follow a fixed longitude is sound but not very responsive without good GPS coverage. – KalleMP Sep 12 '16 at 22:13
  • @IgnacioVazquez-Abrams That's the truth :) I didn't realize the pace of the pole movement had accelerated so much recently. – bitsmack Sep 12 '16 at 23:19
  • 1
    You need to understand the difference between heading and track. This can be very important in boats and airplanes. The GPS alone can never give you heading. Magnetometers can help, but the magnetic field has substantial variation from true north, and the deviation varies from place to place and over time. If you somehow know that your heading and track are lined up (for example in a car on a straight road) then you can use the constant longitude method) – user57037 Sep 13 '16 at 05:01

3 Answers3

2

The accuracy when following lines of longitude via GPS is easy enough to calculate. It's just a question of how much cross-track error you have relative to how long your baseline is (the distance you fly along the line of longitude).

Specifically, suppose you have a GPS receiver with an RMS error of 5m. This is in any direction, so it affects both the cross-track error and the length of your baseline. If you fly a nominal 100m north along a line of longitude as reported by the GPS, you'll have a cross-track error of up to 5m × √2 = 7m. (The √2 is applied because you have error in your start position, and error in your end position, too. These are assumed to be independent.) There will be a length error of up to 7m, too, but we can ignore that since its effect on the final answer is much less.

So, the overall error in your direction is up to 7m for the 100m you traveled, so the error in your concept of due north is \$\sin^{-1}\frac{7}{100}\$, or approximately 0.07 radians = 4.2° RMS.

Since the cross-track error remains constant regardless of the distance traveled, if you double the length of the baseline, the error in the angle will be halved, and so on. If you fly 500m north, then your error will be less than ±1°.

If you use an RTK-capable GPS receiver, the measurement error will be reduced by about two orders of magnitude — centimeters rather than meters. This would allow you to achieve much higher accuracy for a given baseline distance, or to use a shorter baseline for a given level of accuracy.

If your drone/UAV is large enough, you can mount two receivers on it and use a technique known as "short-baseline interferometry" to get an accurate heading measurement directly without any motion at all. The receivers need to be capable of giving you carrier phase measurements, and their antennas need to be on the order of 24" apart. One of my consulting clients has done pioneering work in this area.

Dave Tweed
  • 168,369
  • 17
  • 228
  • 393
0

If you have no metal near drone when calibrating magnetic compass, it should be good for local trips.

If you use accelerometer then stability acceleration and velocity feedback is good in 3 axes but position drift can occur as well as direction during flight, as it integrates acceleration twice.

A MEMs gyro gives the best direction accuracy when calibrated each day of use. GPS position feedback might be possible to correct gyro drift errors during after trip data collection with suitable CPU and firmware.

Since magnetic field lines drift slowly every year and more, closer to the equator, your accuracy increases with latitude,away from equator but not linear.

The North Star and calibration tables is what Surveyor's use to calibrate their equipment with, when I was an student assistant.

Pilots use IFR (radio) and IFR with a map ( I follow railroads)

Tony Stewart EE75
  • 1
  • 3
  • 54
  • 182
0

Single antenna GNSS reception cannot be used for attitude control. It tells you your position in space and time, but nothing about your orientation. Multichannel GNSS receivers with more than one antenna are used for attitude control in avionics and by guided weapons. I am not aware of any consumer grade multi antenna system on the market.

You can however check the orientation of a traveled distance, as David explained in his answer. A current direction may also be available in NMEA output, if the receiver is in motion. This data will typically be less accurate than a long baseline measurement.

If you want to control the direction of your flight, feedback from GNSS will be too slow. I would suggest using a cheap 6 channel IMU and calibrate that from GNSS data.

BTW: You may want to familiarize yourself with the concept of geodetic datum if precision (at high latitudes) matters. Not all the world agrees on what true north actually is. WGS84 is the most common reference frame.

Andreas
  • 1,785
  • 12
  • 18