debi_nav

This package is made for Turtlebot3 waffle_pi navigation in an unknown environment containing both solid and visual obstacles. The package mainly relies on an A* path planning implementation in addition to a low level model predicitve controller for goal tracking. Moreover, the package integrates sensory information by superimposing the traditional map publihsed by gmapping method with another custom map formed by processing raw camera feed in order to create a unified global map. The unified map accounts for the existence of both solid obstacles and visual obstacles with specific RGB features. Accordingly, the robot is able to avoid areas with specific colors and maneuver between walls simultaneously.

System Requirements and Installations

• ROS Noetic
• Python3
• Turtlebot3 simulation package

Dependencies

• Casadi library
• scipy library
• pathfinding library
• OpenCV

Demo Terminal Commands

Terminal tab 1

export TURTLEBOT3_MODEL=waffle_pi

roslaunch debi_nav project.launch

Terminal tab 2

export TURTLEBOT3_MODEL=waffle_pi

roslaunch turtlebot3_slam turtlebot3_slam.launch slam_methods:=gmapping

Terminal tab 3

roslaunch debi_nav debi.launch