I know that accelerometers are used in various devices to detect tilt. Since gravity pulls at a constant rate (1g), the device knows how its orientated relative to the ground depending upon which axises are being “pulled” by gravity.
I also know that accelometers can be used for non-tilt related activities, such as motion detection (for instance a GPS device sometimes use them to infer where your car is when contact with the Satellite has been lost). And since a car doesn’t tilt (well, hopefully it doesn’t!) the accelerometer doesn’t have to rely on gravity to determine where “down” is.
So my question is, can accelometers detect tilt and motion at the same time? For instance, let’s say I took a device that primarly measures tilt, but turned it upside down and then jutted it upward (away from the ground) at the same rate of gravity, 1g. Would the device have any idea how its orientated and that its moving away from the ground?
Nope; in fact, if tilt contributes a sideways vector equal and opposite to the centrifugal acceleration experienced inside a turning vehicle, the accelerometer will respond exactly as if the vehicle were motionless – which is why racecourses, velodromes, and freeway on-ramps are banked.
There’s no reason in principle an accelerometer couldn’t be made to detect both. However, it might have to be a bit more complicated than one designed to measure only one or the other. Tilt, for example, could be measured with nothing but a plumb bob, but a plumb bob won’t be able to distinguish between a car on a tilt and an accelerating car. However, if you replace the string of the plumb bob with a spring, then the spring will be stretched by different amounts in the two cases, so you could tell the difference.
I tried to find a site explaining the “real-time lateral accelerometer” on the new (C6)Corvettes.
All the sites I found only extolled the virtues of its existence.
The description I’ve read implies that it measures both tilt and acceleration.
Are they blowing smoke?
Wait a minute now. I thought that all of the above were banked to increase traction by making the centrifugal force push your car into the road rather than off of it. :dubious:
Could you clarify how this could be done? AFAIK, F=ma=kd=mg=… without much regard to the source of the “F”. I can certainly imagine some clever mechanical configurations making the distinction, but I’m curious how it is done with a swinging pendulum suspended from a spring, in actual practice. Instrumentation design an art I’ve always admired. full of clever tricks
To the OP: a simple accelerometer-only solution would be to (e.g.) rigidly mount several accelerometers on platform with a rigid central mount (e.g. a platform mounted near the center of a car, or sensor distributed to distant corners of a car’s rigid frame). Linear acceleration would have the same magnitude and direction on all the accelerators, but any form of rotation (change in orientation), including “tilt” (roll), “climb” (pitch) or “turn” (yaw) would affect them differently (for, e.g. sensors on opposite sides of the mount). How many accelerometers you need (and their orientations) would depend on how many degrees of freedom you want. Do you care about three axes of linear motion and three of orientation)?
There are many physical arrangements of of gyroscopes and force transducers that will distinguish force from (“global”, “external”) acceleration too, but then you get into a lot of specifics of accelerometer type, design goals, etc. Most accelerometers are just specialized force transducers anyway, and fancier types may use gyroscopes internally.
The amount of (digital or analog) computation you can do matters a lot, too. More data can often be derived from fewer sensors with more computation, but sensor-heavy designs are often more robust, accurate or reliable than ones that rely on more computation.