DPRG
DPRG List  



[DPRG] two wheel balance

Subject: [DPRG] two wheel balance
From: Kipton Moravec kip at kdream.com
Date: Fri Feb 22 23:12:51 CST 2002

I know enough about Kalman Filters to be dangerous.  When it comes to Kalman
Filters I am not a master, I am more of a grasshopper.

It is computationally intensive, and I don't think we can manage with fixed
point math.  And I don't know if we can do it with the slow HC11.  This is a
better roll for your 32-bit processor.

The first step in doing a Kalman is to model mathematically what we are
doing.  We need to model all the forces.  We need the equations for the
torques of the system, and try to figure out the PWM versus torque
relationship, for the voltage you are using. We have to take into account
the inertia, the rotational velocity, the rotational acceleration, friction,
everything.

What we are trying to do is model the system.  Then depending on what we
want it to do, we have state vectors that describe the individual variables
>from the equations.   The values of the vectors can be measured, or derived
>from other measurements.  For example, you can measure position directly if
you have GPS, or you can measure position indirectly based on encoder ticks.

So the first step is write down the equations.

Kip

----- Original Message -----
>From: "David P. Anderson" <dpa at io.isem.smu.edu>
To: <dprglist at dprg.org>
Sent: Friday, February 22, 2002 2:54 PM
Subject: [DPRG] two wheel balance


> Now on to the inertial gravity reference tilt sensor.  Anybody know
> anything about Kalman filters?  :>)
>
> regards,
> dpa
>


More information about the DPRG mailing list