5

Suppose I have a magnetometer. I am on the Earth's surface, away from the influence of other magnetic sources. Assuming the magnetometer is aligned with the reference frame, it should read only the Earth's magnetic field vector at that point.

Suppose now I walk towards a large ferrous object, for example, a large parked truck.

As I approach the truck, what will happen to the magnetic field that I measure? Is the distortion to the magnetic field able to be modelled as additive as a function of position? If not, how could one mathematically describe the distortion? Are there any canonical references on the subject?

EDIT:

In response to the answer(s) below, here are some additional details:

  • Assume that I know what the magnetic field in the local geodetic frame should be to within some desired tolerance, e.g. by using EMM2010 or similar.

  • Ultimately, I do want to calculate attitude, but I have other sensors at my disposal (IMU and GPS) where heading is observable under motion. My desire to to "learn" the magnetic deviation on-the-fly to help stabilise the heading during near-stationary periods. My intuition says that the deviation should be observable when yaw is observable. But I need to start with some sort of model as to what the deviation looks like :)

Glorfindel
  • 1,424
Damien
  • 191

1 Answers1

1

This effect is referred to as magnetic deviation and it used to be difficult problem aboard sailing ships. Metal on board could often cause false compass readings on various headings.

To complicate matters the field is not even around the planet and this effect is known as magnetic declination and this is primarily caused by large iron ore deposits or similar orthographic abnormalities which also change over time (see link).

This effect near metal structures is difficult to model and requires precise physical measurements and 3-d EM modeling to even start and you'll need to include a survey of the area's nominal magnetic field without the object. The field depends on both the object's orientation in the field as well as where you are probing. And likewise anything that can warp a magnetic field can influence what you measure, for example the device itself. The composition of the metal also makes a difference. The combination of all these factors make this a non-linear effect especially with something as large and with random surfaces like a truck. The only real reliable thing you could do is detect if you're close to something deflecting the field or not. So no there are no standard example references for this unless the compensation/measurement/correction of compass deviations would be enough for you.

Edit:

There is no way to really model this non-linear effect. It is also difficult to measure. Usually you need some kind of optical sight line to confirm the distortion you're measuring is due to deviations and not rotation or non-linear-movement. This is even more complicated if you can not keep motion restricted to the same plane (i.e. holding one direction constant either x, y, or z directions).

Imagine measuring the field deviations in 3-d and you'll see how hard this would be to model or measure.

Without metal:

normal

With metal: 3-d field

There are some parallels to trying to measure/correct phase shifts of radio signals due to metal distortions. Back in the '70's Loran navigational systems had problems near large metal structures like the golden gate bridge.

In addition the EMM2010 model is only so helpful as it resolves magnetic anomalies down to 56 km. Anything smaller than 58km could be just localized distortions. The EMM model was compiled from satellite, marine, aeromagnetic and ground magnetic surveys which might not produce an accurate enough model to be of use to you.

If you have a way, like with radar, to detect distance to areas of metal structures you could then autonomously try to figure out it's distortion assuming your GPS positions are accurate enough (they will also experience errors due to metal structures).

I suggest you look at the techniques and tricks of people doing magnetic surveys in the industry: http://www.geosoft.com/media/uploads/resources/technical-papers/Aeromagnetic_Survey_Reeves.pdf

Also don't forget that any device you build will also have it's own built in deviation you need to cancel out with a 3-d model.

user6972
  • 5,778