Hi Haddar,

It would be difficult to use the built-in State-space block directly. The attached slx file is just an extension from Dr.-Ing. habil. Adrian's Simulink model, which is constructed according to the advice from both scientists. You may need to specify your own input signal, u(t).

Alternatively, if you are good at programming. you may consider writing the Time-varying state-space model (in m-file) and then calling it using the S-function block in Simulink.

Good luck.

The only way, that I know, it to build the state space model by using simple blocks, i.e. a clock, multiplicators, multiplexers, etc. (see example)

Hello Ecole, use the Ramp function to generate your t, use the multiplication to get t^2 and multiplication again to get your t^3, multiply each one with the corresponding gain. Then pass all the signal to a multiplexer, your B is time varying vector. Then you can the normal state-space simulink model in matrix form.

I hope this will help you.

Best Regards,

Ashraf

Hi Haddar,

It would be difficult to use the built-in State-space block directly. The attached slx file is just an extension from Dr.-Ing. habil. Adrian's Simulink model, which is constructed according to the advice from both scientists. You may need to specify your own input signal, u(t).

Alternatively, if you are good at programming. you may consider writing the Time-varying state-space model (in m-file) and then calling it using the S-function block in Simulink.

Good luck.

From Matlab's documentation on ss():

"For time-varying, state-space models, specify A as a T-dimensional cell array, where:

Cell t contains an mt-by-mt – 1 matrix.

The matrix in cell t (A{t}) specifies how the states at period t – 1 transition to the new states at period t. "

Now it is not clear neither to me if it is possible to define a continuous TV state space model, since apparently Matlab only lets you define the model at each time interval. A quite clumsy way to handle this, I agree.