Kalman Estimator Structure

[Ahmad Kamal]

Hi,

I want to control the two wheel nxt inverted pendulum robot using state-space controller in Labview.

The state-space model of the robot has

6 states (YawAngle, YawAngluarVel, PitchAngle, PitchAngluarVel, WheelAngle, WheelAngluarVel),

2 inputs (LeftMotrorVoltage, RightMotroVoltage),

3 outputs (WheelAngle, PitchAngluarVel, YawAngle).

I have calculated the full state feedback gains which is a 2x6 matrix.

Since I have 2 control inputs, 3 measured outputs (3 out of 6 states), I need to create an estimator(Kalman).

Could any one can explain me the followings:

1-What would be the size of the kalman G,Q,R matrix?

2-When can I use the elements of G matrix as unity?

Regards,

[Suse]

Hi,

I can give you some links which maybe help you to realize your project:

Kalman Estimator: http://documents.wolfram.com/applications/control/OptimalControlSystemsDesign/10.7.html

How can I learn to program the NXT?: http://www.nxtprograms.com/help/learn.html

NXT Segway with Rider: http://www.nxtprograms.com/NXT2/segway/index.html 

I hope this helps.

Best regards

Suse