MPC Controller with GUI for Racetrack

Overview:

This repository contains a Python implementation of a Model Predictive Controller (MPC) that uses a graphical interface for users to draw a custom racetrack. The controller uses a 2D double integrator model to simulate a car following the user-defined track. The system predicts over a finite horizon using a receding-horizon strategy and visualizes the simulation through an animated Matplotlib plot.

If you are interested in the math and semantics behind the MPC model, check out the THEORY.md file

Features:

• Interactive Track Drawing:
  Users can click on a Matplotlib window to draw a custom racetrack. When done, press Enter to submit the track.

• Smooth Reference Generation:
  The drawn track is processed into a smooth reference trajectory using cubic spline interpolation (via SciPy).

• Model Predictive Control (MPC):
  The MPC controller solves an optimization problem at each time step (using cvxpy) over a finite prediction horizon. It then applies only the first control action and repeats this process in a receding-horizon fashion.

• Real-Time Animation:
  An animated visualization shows a "car" (represented as a rectangle) following the track.

Prerequisites:

Ensure you have Python 3.x installed. The following Python packages are required:

• numpy
• cvxpy
• matplotlib
• scipy

You can install the dependencies using pip:

pip install numpy cvxpy matplotlib scipy

Sometimes cvxpy doesn't install properly when done all together, if it doesn't, then run install it seperately:

pip install cvxpy

Running:

You can run the file with python in the terminal, clone the repository and make sure you are in the root directory. Then run:

python .\MPC.py

Then you will be prompted to draw out your track using points, and then the MPC will simulate a car going around this track. Default MPC-relevant values can be found in the THEORY.md file, and can be altered in the code to change the behavior of the car.

References:

• This GitHub Repository