This repository contains MATLAB and Simulink files developed for implementing perception and control algorithms for unmanned ground vehicles (UGVs), focusing on differential-drive robots. The system uses LIDAR sensor data for mapping, obstacle detection, and navigation control based on Model Predictive Control.
The primary goals of this project are:
- Develop a LIDAR-based perception pipeline for obstacle detection and representation.
- Implement a control strategy for UGVs to follow predefined paths using Model Predictive Control (MPC) and other optimal control methods.
- Simulate and validate the complete system in MATLAB/Simulink and Gazebo.
| Directory/File | Description |
|---|---|
Draw_MPC_point_stabilization.m |
MATLAB script to plot 2D the position of the robot without using the ROS/Gazbeo interface. |
casadi_block.m |
MATLAB script for the MATLAB Function block in Simulink. |
casadi_script.m |
MATLAB script used as base for the casadi_block.m file. |
gera_casadi_function.m |
MATLAB script used to test the control in Gazebo without Simulink. |
shift.m |
MATLAB script used in the gera_casadi_function.m file. |
ros_lidar2d |
MATLAB script used to analyse, filter the data from a 2D LIDAR and plot the obstacles as polytopes. |
ros_lidar3d |
MATLAB script used to analyse, filter the data from a 3D LIDAR and plot the obstacles as polytopes. |
Controltest_gazebo_trajectory_23a.slx |
Simulink models for communicating with Gazebo through ROS. |
MPC_p2p.slx |
Simulink base file for the communication with Gazebo. |
- MATLAB R2023a or later
- Simulink
- Casadi Library - version 3.6.5 or later
- MATLAB Toolboxes:
- MPT3
- Ellipsoidal Toolbox (ET)
- ROS (optional, for real hardware communication or Gazebo simulation)