Update Feb 2, 2008

Attitude Heading Reference System (AHRS)

Feb 2, 2008 PARTS pressentation ahrs_overview

This is a project page for the an open hardware / open source software project to implement a AHRS device for use in a robot kit PARTS is working on.  To do a good job at deadrecogning you need to have a reliable measure of heading at all times.  Simple digital compasses made of 2 axis magnetometers alone don't do a good job.  They are influenced by tilt, and ambient field effects.  Successful ARHS devices go for 1500$, and are made up of a 6-DOF IMU  a 3-axis magnetometer and software.  This effort is to create a bag of parts (with layout) for a successful AHRS along with the software to make it go.  The target application is the Robo-Magellan, but I know there are a number of other applications as well. 

Target BOM for the bag of parts is < 150$.

Prototyping has began.  Currently we have a USB-BitWhacker to talk to some accelerometer sensors and burping the results over the USB channel for processing from a python program. 

We are in a sensor investigation mode and will log test code for interfacing the MCU with the sensors as well as the python code for prototyping the computations.

After we have some experience with the various sensors, we'll settle on the BOM and lay out the PCB.  Then we'll move the algorithms over from python to run on whatever MCU we end up with. 

If you are interested in joining this effort feel free to check out and contribute to the Wiki and other content on this site and send me your thoughts and contributions.  markgross@thegnar.org

We don't yet have a email list, or IRC channel yet but we will as things get going.  (recommendations on what email list server to use welcome, today I'm thinking of using Google or Source Forge for email and code repositories, when it comes to that.)


prototype archive:

  1. Interfacing the USB-BitWhacker with an LIS3LV02DQ:  Python code for the PC, Hacked UBW firmware
    1. The LIS device likes to run at 2-3.3V but you can hit it with 5V if you are lazy.  I think the device suffers in accuracy but it will sort of work.  I have 2 bread boards linking the UBW to the LIS.  See pictures.  pict1, pict2
    2. The PIC hardware SPI fails to work with LIS part, it seems to drive the bits faster than the LIS part can handle at 5V, and when using the voltage splitters, the LIS signal coming back to the SDI pin (RX) is Schmitt triggered and needs a 4.9V logic level to read a one.  Re-writing my code to use a bit banged SPI implementation worked well.
    3. I know there are some nice circuits using diodes and pull ups to connected the LIS part to the PIC pins.  When we are out of the prototype phase we'll need to do a proper interfacing of the 5V, 3.3V, and other parts.
    4. todo: prototype some filtering and integration in python and test on train.
  2. Interfacing the PNI MicroMag 3-axis Magnetometer to the UBW: Python code for the PC, Hacked UBW firmware
    1. todo