NARX for Multi Input Multi Output system Identification

Hello, Can anyone please help how to use NARX method from the Neural Networks Tool box to perform system identification for a system that has Multi inputs Multi Outputs.Thanks.

Answer 1

[Greg Heath]

1. Write and test your code assuming SISO

2. Convert to MIMO and test

3. If you have problems or questions, post relevant code and error messsages

Hope this helps.

Thank you for formally accepting my answer

Greg

[Karima]

Dear Greg, Thanks for answering my question , I'm actually using the NN Toolbox , it's just the first step of defining the input and output vectors that is kind of hard for me. I would be really greatful if you could help me. Knowing that the system that I need to model has 3inputs and 2outputs .

Thanks for your help

[Greg Heath]

The sizes of the input and target matrices are

[ I N ] = size(x) % [ 3 N ]

[ O N ] = size(t) % [ 2 N ]

Answer 2

[Ahmed Hakim]

Hello GUYS

I have the same problem

I have a model of DC motor which has two inputs (the Voltage and the Load torque) and output DC motor speed

Now I would like to identify this DC motor model with its two inputs using neural network to mimic this Motor Model

The idea is that when I change in the load torque as a disturbance the whole output of the DCM being controlled by PID is changed

So I suggested to make online identification for the DC Motor, so that when there is a disturbance change, the Neural network will identify the new MODEL and then a new design for PID will be according to this new model by Neural network

[Ahmed Hakim]

How can I make a system identification using Neural Network for a system with Two Inputs and One output

Answer 3

[fabian]

I need to make the identification of a MISO system with NARX neural network. I have the following code someone can collaborate. thanks

load datoshistoricos.txt;

datoshistoricos;

u=datoshistoricos(:,1:2:3:4)';

y=datoshistoricos(:,5)';

[u,us] = mapminmax(u);

[y,ys] = mapminmax(y);

y = con2seq(y);

u = con2seq(u,587);

d1 = [1:2];

d2 = [1:2];

S1 = 20;

narx_net = narxnet(d1,d2,S1);

narx_net.divideFcn = '';

narx_net.inputs{1}.processFcns = {};

narx_net.inputs{2}.processFcns = {};

narx_net.outputs{2}.processFcns = {};

narx_net.trainParam.min_grad = 1e-10;

[p,Pi,Ai,t] = preparets(narx_net,u,{},y);

narx_net = train(narx_net,p,t,Pi);

narx_net_closed = closeloop(narx_net);

view(narx_net_closed)

mrac_net = feedforwardnet([S1 1 S1]);

mrac_net.layerConnect = [0 1 0 1;1 0 0 0;0 1 0 1;0 0 1 0];

mrac_net.outputs{4}.feedbackMode = 'closed';

mrac_net.layers{2}.transferFcn = 'purelin';

mrac_net.layerWeights{3,4}.delays = 1:2;

mrac_net.layerWeights{3,2}.delays = 1:2;

mrac_net.layerWeights{3,2}.learn = 0;

mrac_net.layerWeights{3,4}.learn = 0;

mrac_net.layerWeights{4,3}.learn = 0;

mrac_net.biases{3}.learn = 0;

mrac_net.biases{4}.learn = 0;

mrac_net.divideFcn = '';

mrac_net.inputs{1}.processFcns = {};

mrac_net.outputs{4}.processFcns = {};

mrac_net.name = 'Model Reference Adaptive Control Network';

mrac_net.layerWeights{1,2}.delays = 1:2;

mrac_net.layerWeights{1,4}.delays = 1:2;

mrac_net.inputWeights{1}.delays = 1:2;