LSTM Replaces SISO MPC

[@JujuhTuberr]

Hello, first of all I want to say that your work is excellent and of great value. I've learned a lot so far. I would like to ask a technical question. Do you think there is any way to develop an algorithm that performs the pid tuning online and automatically? that is, from time to time the system provides the gain constants based on the measurements.

[@apm]

Adaptive control is a well-studied area. Here is a review article: link.springer.com/chapter/10.1007/978-3-642-86420-9_12

[@elyseekoulnodji4340]

Thank you first for the great video, very helpful I have a question : I developed a temperature control system with a specific set point using arduino in a small greenhouse and I logged data in a SD card. Now I wanted to develop an LSTM model that I can implement it as a predictor controller in my greenhouse. my question is to know after training the model and get the prediction accurate, what to do in order to convert that model and use it as a controller? and how to proceed for the implementation? Looking forward to hear from Thank you

[@apm]

Here is an example application of implementing real-time MPC: apmonitor.com/do/index.php/Main/TCLabF Just replace the mpc function call with the lstm function call.

[@haithammohammed9840]

Thank you sir Can I apply the same way for a different plant, for example replacing PID for DC Motor with LSTM?

[@apm]

The approach is general - it should work for any dynamic system where you generate data.

[@roflchopter11]

Can this easily be deployed onto microcontrollers in something like C?

[@apm]

Yes, it appears to be possible www.tensorflow.org/lite/microcontrollers

[@vuhoangdung]

how do you guarantee the stability of the system with LSTM?

[@apm]

There are a couple references such as Chen, S., Wu, Z., Rincon, D. and Christofides, P.D., 2020. Machine learning‐based distributed model predictive control of nonlinear processes. AIChE Journal, 66(11), p.e17013.