> > Hi all!
> > I'm mucking about with an AVR/PIC uC and want to choose a suitable
> > accelerometer
> > (probably from Analog Devices - but please let me know of any suitable
> > alternatives!)
> > to give me good accuracy to calculate 0-62mph times, max lateral G
force,
> braking G's
> > etc
> >
> > What I do need is for someone to point me in the right direction with
> > regards to how
> > I can calculate such things using only the accelerometer and a known car
> > weight?
> >
> > TIA!
> > Adam
>
> Car weight is completely immaterial. This is purely a kinetics problem.
V
= U+AT (Velocity = Initial velocity + acceleration * time. For
> acceleration, you can presume that the initial speed is zero. Simply
> choose a delta t at which to sample the accelaration. Then sum the
> acceleration * delta t. This integrates acceleration over time to give
> instantaneous velocity. At some time this integrated velocity will cross
> the 100 kph boundary. That's your 0-100 kph time. Unfortunately, if
> there's a systematic error to your transducer, you'll integrate that
> systematic error.
>
> Calculating the 100 to 0 time is similar, but you have to take into
account
that you know your end speed, not your initial velocity. Easiest way is
> probably to store the A, t pairs (don't need to store the t if dt is
> regular), then do exacly the same calculation, starting at the first pair
in
a long sequence of 0 or near 0 accleration and assume that is the point at
> which you reached 0 velocity, and working backwards until you get back to
> 100kph.
>
> The explanation's a bit fuzzy, but the technique will work.
>
> One thing to watch is that the longer it takes to accelerate or
decelarate,
the more inaccurate this accelerometer based method will be due to the
> aforementioned integration of systematic errors. This is also why this
> technique is absolutely useless as a velocity measurement technique for
any
reasonable length of time. For instance a .01g error in the transducer
> (approx .1 m/s/s will give an error 1 m/s velocity error - about 3.6 km/h
in
a 10 second 0 to 100 kph time). In general, some sort of direct velocity
> measurement system will be better as the errors do not accumulate.
Personal Navigation In this application, position is determined by dead reckoning (double integration of acceleration over time to determine actual position). Real or fantasy? See answer G.
G. Personal navigation: Dream Land. Long term integration results in the accumulation of error due to small dc errors in the accelerometer, integrator input circuitry, wiring thermocouples, etc. Double integration compounds the errors (t2). Without some way of "resetting" the actual position from time to time, huge errors result. This is analogous to building an op-amp integrator by simply putting a capacitor across it. Even if the accelerometer's accuracy is improved by ten or one hundred times better than currently available, huge errors would still eventually result. It would just take longer to happen. Accelerometers can be used in conjunction with a GPS system when the GPS signals are briefly unavailable. Integration over a short time (a minute or so) can give satisfactory results. Click here for more information about navigation. Go to next question.
My personal favorite is: Car-Noise Cancellation The accelerometer senses low-frequency vibration in the passenger compartment, and the noise-cancellation system nulls it out, using the speakers in the car stereo system. Real or fantasy? See answer J. Car-noise cancellation: Dream Land. While the accelerometer has no trouble picking up the vibration in the passenger compartment, noise cancellation is highly phase-dependent. So while we may cancel the noise in one location (say around the head of the driver), it will probably be increased at other locations.
formatting link
> HTH
>
> --
> Alf Katz
> snipped-for-privacy@remove.the.obvious.ieee.org
>
>