Quickstart on Control System

Table of Contents

 Challenges on controlling a hardware

 Quick introduction to Control System

  How to design it

 Modeling The Plant

 Hardware Requirements

 Wiring Diagram

 Demo Procedure

  (1) Prepare the project

  (2) Run the physical hardware to record input and output

   For estimation

   For validation

  (3) Use System Identification App to estimate plant model

  (4) Simulate and tune

  (5) Test with the hardware

 References

Challenges on controlling a hardware

• How to get started with available hardware on market?
• What to learn and how to quickly implement physically?
• Can I create a controller without knowing much about the "plant"?
• Any fast way of tuning PID parameters?

Quick introduction to Control System

Simple closed-loop control system example:

How to design it

1. Model the plant
2. Tune the controller
3. Test & Verify
4. Deploy

Modeling The Plant

The problem is I don't know the specification of the motor and its driver (the plant)

So, this is where we need System Identification

Hardware Requirements

• Arduino Uno R3
• Type-B USB Cable
• L293D
• 12VDC brushed motor (or similar)
• 12V power supply
• 12V barrel jack power socket

Wiring Diagram

Demo Procedure

(1) Prepare the project

open system_identification_arduino.prj

(2) Run the physical hardware to record input and output

load_system("record_plant");
set_param("record_plant",SimulationMode="external");

For estimation

start_recording("record_plant");
estimation_iddata = recording_output(yout);
save('Output\estimation.mat', 'estimation_iddata');
plot(estimation_iddata)

For validation

start_recording("record_plant");
validation_iddata = recording_output(yout);
save('Output\validation.mat', 'validation_iddata');
plot(validation_iddata)

(3) Use System Identification App to estimate plant model

Use these object names and click import for each of them.

1. estimation_iddata
2. validation_iddata

Then click "Close".

You will see "estimation_idddata" and "validation_iddata" in Data Views. Drag and drop accordingly to this illustration.

Configure exactly as below. Feel free to modify if you know what you're doing.

Then click "Estimate" and wait.

Once done, drag the result as below to workspace.

Done! Now you have "plant_model" variable in Workspace, which represents your motor and tachometer as a plant.

If you want instant result, here's the custom function:

plant_model = identify_the_plant(estimation_iddata, validation_iddata);

(4) Simulate and tune

open tune_controller.slx

Double click the PID block

Make sure changes applied and tuning method is "Transfer Function Based", then click Tune button

Adjust these sliders below then hit "Update Block" button.

Click OK on Block Parameter dialog. Your PID parameters (Kp, Ki, Kd, and Kn) are now synced to other models via "Data\tuning_parameters.sldd".

(5) Test with the hardware

open hardware_in_the_loop.slx

References

• Get Started with System Identification Toolbox