Simulation system for the development of autonomous vacuum cleaner navigation

ROS 2 - based project implementing simple covering area with turtlebot3 as vacuum cleaner.

Simulation system for autonomous mobile robot navigation evaluation, which enables to create dynamic simulation environment with random rooms and emerging obstacles. System is described in Domain Specification Language SPSysML (https://arxiv.org/abs/2303.09565) , which rates system simulation-physical integrality. The thesis proposes indicators used to examinate navigation algorithms. The work is a proposal of a standard for quality indicators used when evaluating navigation algorithms of mobile robots

Installation

ROS 2 Foxy https://docs.ros.org/en/foxy/Installation.html

TurtleBot3 https://emanual.robotis.com/docs/en/platform/turtlebot3/quick-start/

Install dependencies

pip3 install meshlib
pip3 install aspose-words

Clone project

cd ~/vc_ws/src
ros2 pkg create --build-type ament_cmake vacuum_cleaner_pkg
sudo rm -r vacuum_cleaner_pkg
git clone https://github.com/MaciejGroszyk/vacuum_cleaner_pkg.git
cd ..
colcon build

Run

cd vc_ws/ && source ~/ros2_foxy/install/setup.bash && source ~/turtlebot3_ws/install/setup.bash && source install/setup.bash

export TURTLEBOT3_MODEL=waffle && ros2 launch turtlebot3_gazebo empty_world.launch.py

Manualy spawn map from /home/user/.gazebo/models and drop robot to map frames and run algorithm.

-- random walk
ros2 run vacuum_cleaner_pkg main_node random
-- spiral walk
ros2 run vacuum_cleaner_pkg main_node spiral
-- snaking walk
ros2 run vacuum_cleaner_pkg main_node snaking

Generate world with command

~/vc_ws/src/vacuum_cleaner_pkg/scripts
-- generate model 3D
pip3 worldMapModelGenerator.py
-- generate model 2D
pip3 worldMapPngGenerator.py
-- spawn obstacles
pip3 gazeboRandomSpawnModel.py

Requirement specification

Project structure

World synchronization

Flow diagram

Load visualization

Load visualization example

Generating the real world to simulation

Generating the real world to simulation example

Generate random visualization

Generate random visualization example

Spawninng obstacles

DT/PT decomposition – Mirroring Agent Groups specification

Simulation agent

Physical agent

Agent decomposition

Simulation agent decomposition

Physical agent decomposition

Structure evaluation

Controller integrity factor IFF = 1. The higher the IIF value, the more software components are shared between the simulation and physical embodiments. Maximum IIF value is 1 - all abstract hardware parts of the system are common.

Driver generalisation factor DGF = 1. The higher the value, the easier it is to integrate physical and simulated hardware in the future.

Digital Twin coverage DTC = 1. All physical parts have digital twins.

Mirror integrity factor MIF_Robot = 1. Parts of the system between the physical and digital twins that make up the robot are shared.