Using fixed FPGA capacity for a variable number of Simulink signal channels

In Simulink, I am modelling a MIMO system. One section involves a large set of impulse responses : P inputs, Q outputs, and all possible (P -> Q pair) impulse responses are known. Normally, I could simply implement P x Q FIR blocks, combine their outputs appropriately and everything would be fine.

However, there are two complications :

1. I need to run the model, then modify P and/or Q, then run the model again. Multiple times, changing P and/or Q each time. And the process of successively running multiple experiments needs to be automated.
2. The model will ultimately run on an outboard FPGA-based device (Speedgoat / dSpace / etc) with limited processing capacity. It is therefore important to avoid unnecessary calculations. Otherwise I could simply implement the maximum number of channels, and fill some of them with zeros if they are not being used during a particular experiment.

I do not know how to deal with both of these problems at once, and have not managed to find any clues online.

A similar problem occurs elsewhere in the same model, with vector-array addition blocks in which the numbers of inputs and outputs need to be variable.

Thank you for any advice !

Answer 1

[Sahas]

Hi @Ruairidh,

As per my understanding, you want to model a MIMO system in Simulink with dynamically changing inputs P and Q automate the process of conducting multiple simulations. Ultimately, this model will run on an FPGA-device with limited processing capacity.

The steps and MathWorks documentation links provided below will help you get started with the task at hand:

• I would suggest to parameterize P and Q and define them as variables to allow dynamic adjustments in successive experiments.
• Use "MATLAB Function blocks" for changing the values of P and Q in between experiments as required. Alternatively, use PreLoadFcn or InitFcn model callbacks for the same.

https://www.mathworks.com/help/simulink/ug/what-is-a-matlab-function-block.html

https://www.mathworks.com/help/simulink/ug/model-callbacks.html

• For automating the experiments, develop a MATLAB script for automatically setting model parameters, running simulations and handling outputs. Use MATLAB's sim and set_param functions to control Simulink model from MATLAB editor.

https://www.mathworks.com/help/simulink/slref/sim.html

https://in.mathworks.com/help/simulink/slref/set_param.html

• For optimizing the code, use conditional execution subsystems such as "Enabled Subsystems", "Switch Blocks" to route signals and minimize unnecessary calculations.
• To deploy the model on an FPGA device, use HDL Coder Workflow Advisor.

https://www.mathworks.com/help/hdlcoder/ug/using-the-hdl-workflow-advisor-window.html

Hope this is beneficial!

[Ruairidh]

Good morning @Sahas, and thank you kindly for the reading list. I will explore, test, and report back...

Thanks !

Answer 2

[Kiran Kintali]

Have you considered tunable parameter usage in HDL Coder?

https://www.mathworks.com/help/hdlcoder/ug/generate-code-for-tunable-parameters.html

Please do not hesitate to reach to tech support for the limited processing capablity question.

[Ruairidh]

Thank you kindly @Kiran Kintali. I was unaware of this and will look into it.

Go well !