I have used accelerometers based on microelectromechanical systems (MEMS) for both hobbyist projects and more legitimate designs. In the latter category, I used one as an inclinometer. As the name suggests, I was measuring the angle of inclination (more properly, the deviation from the perpendicular) as part of a CT scanner gantry. We needed to know the angle the gantry was tilted before we shot X-ray beams through the patient.
The accelerometer provided absolute position information. We also used a potentiometer coupled to the gantry's tilt axle, so that we had a redundant method (using a different technology) to measure the tilt angle. In medical diagnostic equipment, it is good design practice (i.e., essential and mandatory) to use redundant systems to monitor certain vital parameters. Any subsystem that moves the patient or moves equipment around the patient — and could cause injury to the patient or artifacts in the imaging — must be carefully managed.
A MEMS accelerometer contains a micromachined beam that deflects when external forces act on it. The deflection can be measured by different means, but measuring the capacitance between the beam and nearby electrodes is a common method. The external forces are either gravity (the inclinometer application) or acceleration acting on the beam's inertial mass (f=ma).
The hobby project to which I referred also used an ADI device, but this time, it actually measured acceleration. I used it in an automotive application to monitor and display positive and negative acceleration. The display itself was amusing. However, there is more to that than room here will allow. I'll leave that for another blog.
If you combine a three-axis accelerometer with a three-axis magnetometer (functionally a compass), you can create a reasonably good subsystem to track motion and, by extension, location. Of course, it helps if you know where you are to start with, and it helps even more if you can double-check where you are every so often in some absolute sense.
Researchers at the University of Amsterdam have teamed up with STMicroelectronics to develop a bird-tracking system that does pretty much what I described above. They are using an IC that is a three-axis accelerometer and a three-axis magnetometer, all in a 28-pin, 5mm-by-5mm land grid array (LGA) package.
Their system also uses a GPS positioning subsystem that checks absolute location every three seconds, a data logging subsystem, temperature sensors, and a lithium battery plus solar cell to charge it. This all weighs about as much as a 25-cent piece, so it can actually be attached to a bird — probably a small hawk, not a house finch.
Knowing the bird's absolute position every three seconds — and using the accelerometer and magnetometer to fill in data in the interim — the researchers can log the bird's flight accurately. Even the bird's orientation (pitch, tilt, and yaw) can be tracked, so the scientists can tell how it is dealing with crosswinds.
The information is used to verify computer models for bird behavior and to analyze migration patterns. By extension, the researchers can draw conclusions regarding the environment and climate change.