如何将离线训练好的LSTM网络，放在simulink中实时预测

simulink中的Stateful Predict模块可实现此功能。

提示：matlab2021版本才有此模块！！！

---

Model Example

1、 Load Pretrained Network

load JapaneseVowelsNet;

2、 Load Test Data

[XTest,YTest] = japaneseVowelsTestData;
X = XTest{94};
numTimeSteps = size(X,2);

3、 Simulink Model for Predicting Responses

open_system('StatefulPredictExample');

4、 Configure Model for Simulation

set_param('StatefulPredictExample/Input','Value','X');
set_param('StatefulPredictExample/Index','uplimit','numTimeSteps-1');
set_param('StatefulPredictExample/Stateful Predict','NetworkFilePath','JapaneseVowelsNet.mat');
set_param('StatefulPredictExample', 'SimulationMode', 'Normal');

5、 Run the Simulation

out = sim('StatefulPredictExample');

6、 Plot the prediction scores

scores = squeeze(out.yPred.Data(:,:,1:numTimeSteps));

classNames = string(net.Layers(end).Classes);
figure
lines = plot(scores');
xlim([1 numTimeSteps])
legend("Class " + classNames,'Location','northwest')
xlabel("Time Step")
ylabel("Score")
title("Prediction Scores Over Time Steps")

trueLabel = YTest(94);
lines(trueLabel).LineWidth = 3;

---

matlab官方文档：https://www.mathworks.com/help/releases/R2021a/deeplearning/ug/predict-update-network-simulink.html