Diploma thesis

---

This repository contains ROS based software part of the project.

The aim of the project is to create robot which can be operated manually or drive autonomously. Robot should be able to navigate to charging station, when it needs to, charge itself and procced to normal state completely without external supervision. Fiducials in environment are used for localization and for augmented reality.

---

Description of robot:

Sensors:

• RPi camera
• LIDAR - only for SLAM
• IMU
• 4 Wheel encoders

Hardware:

• Raspberry Pi 3B+
• Custom PCB with Arduino SAMD Zero
• Working station PC - Ubuntu

Mechanics:

• 4 Wheel Skid Steer
• Front servo gripper

---

Localization using fiducials:

Use of ArUco for localization in environment

---

Augmented reality using fiducials:

Hosted on another repo.

---

Simulation - Created simulation environment in Gazebo, URDF robot model

Working features in Gazebo:

• SLAM
• Localization using fiducials
• Navigation - use of move_base_flex
• State machine for navigation and part of charging cycle

Real robot - Currently testing procedures in laboratory

Working features in real world:

• SLAM
• Localization using fiducias
• Navigation - use of move_base_flex
• Following object with ArUco marker
• State machine for navigation and complete charging cycle
• Automatic comparsion of map generated by SLAM with ground truth map

---

State machine:

Complete SMACH:

Navigation loop:

Navigation to charging plug:

Charging:

Automatic comparsion of created map from SLAM with ground truth map