Autonomous Indoor Delivery Robot Using ROS-Based Navigation on Kobuki QBot

CS3340: Robotics and Automation Final Project

Department of Computer Science and Engineering

---

👥 Team: RoboSoulz

Name	Index Number
Birajith K.	230091H
Rafi M.A.A	230505J
Saran S.	230582N

---

📌 Project Overview

This project focuses on the design and implementation of an Autonomous Indoor Delivery Robot using the Kobuki QBot platform integrated with a Kinect RGB-D sensor.

The robot is designed to navigate autonomously within an indoor environment and deliver items between predefined locations. The system integrates core robotics components—perception, planning, and control—within a ROS 2-based architecture.

---

🎯 Objectives

• Generate an occupancy grid map using depth sensor data
• Implement path planning algorithms (A*, Dijkstra)
• Enable autonomous navigation between predefined delivery points
• Integrate perception, planning, and control modules
• Demonstrate a complete indoor delivery task

---

🏗️ System Architecture

The system is divided into the following modules:

🔹 Perception Module

• Acquire RGB and depth data from Kinect sensor
• Generate occupancy grid map
• Detect obstacles in the environment

🔹 Planning Module

• Select target delivery locations
• Generate collision-free paths using planning algorithms
• Handle path updates if required

🔹 Control Module

• Convert planned paths into velocity commands
• Control robot movement using ROS cmd_vel
• Maintain stable trajectory tracking

🔹 ROS Integration

• Implement each module as ROS 2 nodes
• Communicate via ROS topics
• Visualize system using RViz

---

📁 Repository Structure

final-project-robosoulz/
├── docs/
│ ├── images/
│ ├── logs/
│ │ └── week08/
│ │     └── local_setup_verification.txt
│   ├── setup_notes/
│   │   ├── week08/
│   │       ├── week08_setup.md
│   │       └── week08_node_summary.md
│   │   ├── week09/
│ ├── weekly_reports/
│ │     └── week08_report.md
│
├── src/
│ ├── qbot_navigation/     # Main project package (perception, planning, control)
│ ├── qbot_description/    # Robot model and visualization files
│ ├── interfaces/          # Custom ROS interfaces (if used)
│
├── README.md

---

⚙️ Technologies Used

• ROS 2 (Jazzy)
• Python (ROS Nodes)
• Kobuki QBot
• Kinect RGB-D Sensor
• SLAM / Mapping Tools
• RViz for visualization

---

🚀 Setup Instructions

1. Clone the Repository

git clone <your-repo-link>
cd final-project-robosoulz

2. Build

colcon build

3. Source

source install/setup.bash

4. Run

ros2 run qbot_navigation odom_node

---

📅 Project Timeline

Week	Task
Week 08	Setup and ROS environment verification
Week 09	Perception (Kinect + mapping)
Week 10	Path planning
Week 11	Motion control
Week 12	System integration
Week 13	Testing and documentation

---

🧪 Week 08 Progress

✅ Completed

• Repository initialized and structured
• ROS 2 environment verified (Jazzy)
• Workspace successfully built using colcon build
• Integrated packages from previous labs
• Renamed package to qbot_navigation
• Verified package detection in ROS
• Verified executables:
  • mapper_node
  • navigation_server
  • odom_node
  • qbot_controller
• Successfully launched ROS nodes locally
• Verified node registration (/map_node)

❌ Pending

• Raspberry Pi setup
• QBot hardware connection
• Kinect sensor integration
• Hardware communication testing

---

📚 References

• Kobuki QBot Documentation
  https://kobuki.readthedocs.io/en/devel/index.html

• Kobuki Python Interface
  https://github.com/IntellisenseLab/kobuki-python

• ROS 2 Navigation Stack (Nav2)
  https://docs.nav2.org/

• Thrun, S., Burgard, W., & Fox, D. (2005)
  Probabilistic Robotics, MIT Press

---

📌 Notes

This project builds upon ROS-based lab implementations. Current progress includes software setup and validation on a local machine. Hardware integration will be performed in subsequent phases.

---

Project Status

🟡 In Progress — Week 08 Completed (Software Setup Stage)