7 - Accelerometers
Published online by Cambridge University Press: 03 May 2011
Summary
General measurement objectives
Accelerometers convert accelerations into deflections or stress deviations. The deflections or stresses are transformed into an electrical vector signal Ȳout that should represent a more or less optimal estimate of the acceleration vector. (Estimates of a function f(t) are denoted by.) f(t)The most common estimation algorithms are simple linear filtering procedures that smooth the noise of the reactions captured.
From a more general viewpoint, an inertial sensor is a system, the output signal of which depends on six inertial forces caused by three linear accelerations and three rate signals or their derivations:(ax, ay, az, Ωx, Ωy, Ωz). Even for sensors, which usually are approximated by 1D systems, the impact of all inertial forces is virtually present in the form of cross-coupling effects. In a 1D system the response of the five undesired outputs is suppressed to negligible levels in comparison with the reaction of the main component.
Ideally, a one- to three-dimensional accelerometer should deliver an output signal Ȳout that satisfies the following criteria.
Within a given measurement range it represents a linearly scaled estimate of the input signal,
where S is the scale-factor matrix. is the offset, which often is called bias.
It constitutes a “good” estimate of the input component (for instance in terms of the minimum squared error or of the maximum a-posteriori probability), i.e. resolves it with high accuracy within the specified measurement range and bandwidth. The resolution is usually characterized by the spectral density of the output signal in that is equal to the noise density and, thus, that can be differentiated from it.
[…]
- Type
- Chapter
- Information
- Inertial MEMSPrinciples and Practice, pp. 283 - 363Publisher: Cambridge University PressPrint publication year: 2011
- 4
- Cited by