README.md

Online-LQR

This repository is dedicated to the implementation and study of Online Linear Quadratic Regulator (LQR), which combines traditional LQR with reinforcement learning techniques.

Contents

• MATLAB Files:

  • CTLQR_PI_3x3_Sogang.m: Continuous-time LQR with Proportional-Integral control.
  • DTLQR_Q_3x3_Sogang.m: Discrete-time LQR with state-space representation.
  • DTLQR_Q_3x3_Sogang_Nonlin_ref.m: Discrete-time LQR with nonlinear reference.
  • DTLQR_Q_3x3_Sogang_Nonlin_ref_PI.m: Discrete-time LQR with nonlinear reference and PI control.

• Python Files:

  • MPC_simple_pendulum.py: Model Predictive Control (MPC) implementation for a simple pendulum system.

Getting Started

Prerequisites

• MATLAB (for running the .m files)
• Python 3.6 or higher (for running the .py files)
• Required Python libraries: numpy, scipy, matplotlib

Installation

1. Clone the repository:

   git clone https://github.com/mincasurong/Online-LQR.git
   cd Online-LQR

2. Install the required Python libraries:

   pip install numpy scipy matplotlib

Usage

Running MATLAB Files

Open the .m files in MATLAB and run them to see the results of various LQR implementations.

Running Python Files

To run the Model Predictive Control example for a simple pendulum:

python MPC_simple_pendulum.py