NN-based Control Methods for Dynamic Systems

Reference Tracking Neural Network Controller for Dynamic Systems

The controller utilizes Neural Networks to control a nonlinear dynamic system by tracking a given reference signal. The main goal is to minimize the error between the system output and the desired reference trajectory.

Block Diagram

Components Description

1. Reference Model:

   • The reference model is defined by the transfer function: $$G_m(s) = \frac{K}{\frac{1}{\omega_n^2}s^2 + \frac{2\xi}{\omega_n}s + 1}$$
   • Generates the desired reference signal.
   • Uses the reference input which is given by: $$r(k) = \sin\left(\frac{2\pi k}{25}\right) + \sin\left(\frac{2\pi k}{10}\right)$$

2. Dynamic System:

   • The nonlinear dynamic system is represented by: $$y(k+1) = \frac{y(k) y(k-1) u(k) + u^3(k) + 0.5 y(k-1)}{1 + y^2(k) + y^2(k-1)}$$

3. NN Controller:

   • Adjusts the control input to minimize the tracking error.
   • Utilizes gradients and parameters of the RBF NN model to update the control signal.

Result

Reference

[1] Slema, S., Errachdi, A., & Benrejeb, M. (2018, March). A radial basis function neural network model reference adaptive controller for nonlinear systems. In 2018 15th International Multi-Conference on Systems, Signals & Devices (SSD) (pp. 958-964). IEEE.

Actor-Critic-Identifier DRL Framework for Pendulum Balancing Problem