NN Predictive Controller for MIMO systems

Good morning
I'm trying to realize a NN predictive controller for a system (which I only have as a neural network) with more than one manipulated value for a project at the uni.
The Problem is, whenever I try to do the system identification with the data of my planet it gives me the error message "The input and output arrays mut be row (1*N) or column (N*1) variables".
And I'm stuck right there, so could someone help me?
Is there any way to implement the NN predictive controller for MIMO or MISO systems?
many thanks in advance

6 Comments

I would like to know about it too.
Hi @Nohayla Ajmi, do you have the mathematical model of your MIMO system?
hello @Sam Chak, No there is no mathematical or physical model of the system I would like to model it using measurement data. I have been able to build predictive models using muti-inputs and I would like to make a control law ( Otherwise if this is not possible I would like to at least know and prove if the system is controlable/observable or not)
Sam Chak
Sam Chak on 2 Aug 2022
Edited: Sam Chak on 2 Aug 2022
@Nohayla Ajmi, thanks for your reply. If your ultimate goal is to a design a control law, perhaps you can try to estimate the model using some techniques in the System Identification Toolbox.
@Abdulhamid Jalab did you know how to implement the nn predictive controller with MIMO system, please if yes let me know how..

Sign in to comment.

Answers (1)

Manikanta Aditya
Manikanta Aditya on 11 Jan 2024
Edited: Manikanta Aditya on 11 Jan 2024
Hi Abdulhamid,
Looks like you are trying to realize a NN predictive controller for a system with more than one manipulated value for a project. Then you are a error and also you would like if there is a way to implement the NN predictive controller for MIMO or MISO systems.
Regarding the error ‘The input and output arrays mut be row (1*N) or column (N*1) variables’, the issue seems to be with the dimensions of the input and output arrays. When working with matrices or arrays, you need to ensure that the dimensions match the expected format. You might need to transpose your array to match the expected format.
Model Predictive Control (MPC) is a method that can be used for controlling systems. MPC uses a model of the system to predict future behavior and optimize the process over a defined time period.
Please refer to the following reference to know more about
Hope it helps!

Categories

Find more on Model Predictive Control Toolbox in Help Center and File Exchange

Products

Asked:

on 27 Jul 2022

Edited:

on 11 Jan 2024

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!