I would like to design a MLPNN controller for Simulink but I don't exactly know what the input(s) and output(s) of my neural network controller will be? And how can I enter these inputs to neural network MATLAB code? Is there any MATLAB code for entering these inputs to MATLAB code? Or do I need some data-sets for my inputs? If so, how can I find the data-sets? I also don't want to use MATLAB toolbox. Your answers will be appreciated and help me a lot.
best wishes
chrome-extension://oemmndcbldboiebfnladdacbdfmadadm/http://control.ee.ethz.ch/~apnoco/Lectures2009/10-ANN.pdf
Example to train and simulate MLP in MATLAB and Simulink:
Training data set (MLP = one input layer (with one input), one hidden layer and output layer (with one output)):
x = [-1.000000E+00 -9.500000E-01 -9.000000E-01 -8.500000E-01 -8.000000E-01 -7.500000E-01 -7.000000E-01 -6.500000E-01 -6.000000E-01 -5.500000E-01 -5.000000E-01 -4.500000E-01 -4.000000E-01 -3.500000E-01 -3.000000E-01 -2.500000E-01 -2.000000E-01 -1.500000E-01 -1.000000E-01 -5.000000E-02 3.191891E-16 5.000000E-02 1.000000E-01 1.500000E-01 2.000000E-01 2.500000E-01 3.000000E-01 3.500000E-01 4.000000E-01 4.500000E-01 5.000000E-01 5.500000E-01 6.000000E-01 6.500000E-01 7.000000E-01 7.500000E-01 8.000000E-01 8.500000E-01 9.000000E-01 9.500000E-01 1.000000E+00];
y= [8.996000E-01 7.903000E-01 6.810000E-01 5.717000E-01 4.624000E-01 3.531000E-01 2.438000E-01 1.345000E-01 2.520000E-02 -8.410000E-02 -1.934000E-01 -3.027000E-01 -4.120000E-01 -5.213000E-01 -6.306000E-01 -7.399000E-01 -8.492000E-01 -6.369000E-01 -4.246000E-01 -2.123000E-01 1.355186E-15 2.070050E-01 3.847503E-01 5.058764E-01 5.481037E-01 4.996742E-01 3.640109E-01 1.621403E-01 -6.851182E-02 -2.797940E-01 -4.227630E-01 -4.602351E-01 -3.788687E-01 -1.967830E-01 3.705930E-02 2.537087E-01 3.841045E-01 3.828430E-01 2.475974E-01 2.535759E-02 -2.001342E-01];
net = newff(x,y,10); % ceation of the MLP with 10 hidden neurons (one hidden layer)
net.trainParam.epochs = 1000; % number of iterations
net.trainParam.goal = 0.00001; % MSE
net.trainParam.lr=0.1 ; % training parameter ette
[net,tr] = train(net,x,y); % training using LM algorithm
plotperform(tr) ; % Display different errors
r = sim(net,x) ; % Simulation of the obtained MLP
plot(x,y,’black-’,x,r,’black+’); %Plot desired output y and obtained output r of the MLP
We can use the Neural Network Toolbox/nntool to train the MLP. We can also use Simulink to simulate the obtained MLP.
Below link gives your the comprehensive details of the requested topic
http://www.academia.edu/7064142/NEURAL_NETWORKS_Basics_using_MATLAB_Neural_Network_Toolbox
Use NNTOOL in Matlab for designing and simulating your owm MLP NN.
It is GUI based toolbox very easy to understand. If you require any help please feel free to ask.
- Badges
- Science topic