Internal Memo:
Support guide for end users with IMU and require ENU estimates
Date:
16 July, 2014
Author:
Torstein Pedersen
Microstrain IMU coordinates
Nortek coordinates
16 cm
Y
Z
X
Figure 1 Microstrain IMU mounted on the Vector circuit board. Definition of the different coordinate systems indicated.
Microstrain – Vector Integration
The Vector end user has the option to integrate a Microstrain IMU sensor which is intended to help
understand how the Vector’s own motion influences the measurements. This is clearly a special
case that requires that the end user has competence and understanding of the data products
output from the IMU and recorded by the Vector.
The Microstrain IMU is incorporated in the Vector by simply using the magnetometer contact
interface. This means the standard tilt and compass are removed and replaced with the IMU. This
means there is no standard Nortek compass system with the IMU integrated. The IMU is not
intended to be used as an internal compass, such that the user may collect data in ENU coordinates.
The IMU requires that the Vector collects data in either XYZ or beam coordinates. Conversions
beyond this (into ENU) may be done in post processing by using information from the IMU.
The center of the Microstrain is approximately 16 cm from the end of the endbell where the probe
stem is affixed.
A complete understanding and usage of the Microstrain IMU requires a review of the 3DM-GX3-25OEM product page: http://www.microstrain.com/inertial/3DM-GX3-25-OEM.
Accessing the IMU
Although it is not necessary the end user may access the Microstrain IMU by setting the Vector in a
transparent mode. The steps are to
1)
2)
3)
4)
Send a break to establish standard contact with the Vector via the Terminal Emulator
Send [54 53 00 00 hex] to set in transparent mode (see Figure 2)
Close the Vector software so that the comm port is available for the Microstrain software.
Connect to the IMU with the Microstrain Monitor software.
NOTE The software assumes that the IMU is oriented in the Micostrain frame of reference in order
to use the X, Y, Z, tilt, roll, and heading quantities sensibly.
Figure 2 Commands in Terminal window that sets Vector in transparent mode and allows communication to the Microstrain
IMU.
The IMU maintains the default factory setting, as indicated by dialog box of Figure 3. The end user
may change these settings with the Microstrain interface software; note however that Nortek does
not take responsibility for errors that may occur if changes are made beyond the factory settings.
Figure 3 Window in Microstrain software that shows IMU's device settings.
Data Products
When using the IMU sensor, the end use is required to record in XYZ or Beam coordinates. The user
is permitted to output (and record internally on the Vector) IMU data in three formats:
1) Accelerometer, Angular Rate, Magnetometer, and Coordinate Transfer Matrix.
2) Accelerometer, Angular Rate, and Magnetometer.
3) Delta Angle, Delta Velocity, and Magnetometer.
The option which includes the coordinate transfer matrix is the output recording option that allows
the user to estimate the orientation of the instruments as well as estimate ENU quantities.
Coordinate Transformation
The Vector and the IMU have different definitions of the XYZ and East-North-Up axes. This means
that there is a little “book keeping” in order to use the Microstrain coordinate transform matrix to
estimate ENU velocity estimates in the Nortek Vector defined coordinates.
Figure 1 shows the Microstrain IMU sensor mounted on the Vector circuit board and the definition
of the axis. Further to this, Nortek’s order of axis for the Earth frame of reference is East-North-Up,
while Microstrain uses a North-East-Down convention.
Below is a coordinate conversion to arrive at velocity in ENU coordinates using the recorded
transformation matrix (M). The transformation matrix is a 3x3 matrix and is found in the *.DAT file
(see the header for specification).
The conversion of local to earth coordinates for Microstrain space is
𝑉𝑙𝑜𝑐𝑎𝑙 = 𝑀 ∙ 𝑉𝐸𝑎𝑟𝑡ℎ
Again M is the transform matrix recorded by the Vector.
The transformation between the Vector and the Microstrains Earth coordinate definition is as
follows
𝑉𝐸𝑎𝑟𝑡ℎ 𝑁𝑜𝑟𝑒𝑘 = 𝑇𝐸𝑎𝑟𝑡ℎ ∙ 𝑉𝐸𝑎𝑟𝑡ℎ 𝑀𝑖𝑐𝑟𝑜𝑠𝑡𝑟𝑎𝑖𝑛
In a similar manner we can convert between relative local coordinates between the Nortek and
Microstrain coordinate definitions as
𝑉𝐿𝑜𝑐𝑎𝑙 𝑁𝑜𝑟𝑡𝑒𝑘 = 𝑇𝐿𝑜𝑐𝑎𝑙 ∙ 𝑉𝐿𝑜𝑐𝑎𝑙 𝑀𝑖𝑐𝑟𝑜𝑠𝑡𝑟𝑎𝑖𝑛
Where the transformation matrix are defined as follows,
0 1
𝑇𝐸𝑎𝑟𝑡ℎ = [ 1 0
0 0
0
0 ]
−1
0 0
𝑇𝐿𝑜𝑐𝑎𝑙 = [0 1
1 0
−1
0]
0
The final result for converting from the Vectors XYZ velocity to its ENU estimates is as follows
−1
𝑉𝐸𝑎𝑟𝑡ℎ 𝑁𝑜𝑟𝑒𝑘 = 𝑇𝐸𝑎𝑟𝑡ℎ ∙ 𝑀 −1 ∙ 𝑇𝐿𝑜𝑐𝑎𝑙
∙ 𝑉𝐸𝑎𝑟𝑡ℎ 𝑀𝑖𝑐𝑟𝑜𝑠𝑡𝑟𝑎𝑖𝑛