Autonomous Robotics Lab - Team Zirbi

• Autonomous Robotics Lab - Team Zirbi
  • Documentation
  • Getting Started
    • Setup
    • Simulation startup
  • Architecture Overview
    • Bringup & Integration
    • Core Infrastructure
    • Perception & Vision
    • Knowledge & State Management
    • Decision Making & Behavior
    • Manipulation
    • Movement & Navigation
    • Speech & Audio
    • Safety
    • Visualization & UI
  • Contributing
  • Authors and acknowledgment

This is the code repository for the home-assistant robot of team Zirbi developed in Autonomous Robotics Lab

This Project aims to develop an fully automatic @Home League robot for the RoboCup.

Goals of this Robot are:

• be a social and cooperative robot for everyday use
• accomplish easy tasks in home environment
• learn new tasks from visibel demonstration or verbal instructions
• bring joy to users and developers

Documentation

The doc folder contains the general project documentation. All code documentation can be found in the respective source code files.

Getting Started

Setup

This project needs to be cloned as part of the arlab_docker project in order to work.

# Do not clone arlab manually, just clone arlab_docker
git clone https://git.rz.uni-augsburg.de/imech-m/autonomous_robotics_lab/arlab_docker
cd arlab_docker

# Run the automated setup script (Clones this repo and sets up the environment)
./user_setup.sh

For further information about the setup, dependencies and the containers in general, look at the arlab_docker README.

Simulation startup

To start the simulation environment, first ensure the Docker container is running:

# Navigate to the arlab_docker root
cd /path/to/arlab_docker
# Build and start the Docker container (choose your GPU type)
docker compose -f docker-compose.dev.<your-gpu-type>.yml up --build

Available GPU types / platforms:

• cuda - NVIDIA GPU support (Linux/WSL Lab PCs)
• intel - Intel GPU (Vulkan)
• rocm - AMD GPU (ROCm)
• wsl - Windows WSL2 with WSLG GPU support

Once the container is running, enter it and launch the simulation as described in the arlab_bringup documentation.

# Find container name
docker ps
# Enter the container (Or just attach Vs Code with the Container Tools extension)
docker exec -it <container-name> bash

# Start basic simulation
ros2 launch manipulator_description manipulator.full.sim.launch.py

Architecture Overview

The following tables provide an overview of the ROS 2 packages in the code directory, organized by functional category.

Bringup & Integration

Package	Description	Documentation
arlab_bringup	Main launch package for the robot system	Read more

Core Infrastructure

Package	Description
arlab_asyncio_executor	Async executor for running Python nodes in ROS 2
arlab_common	Common utilities and shared code
arlab_common_interfaces	Common ROS 2 action and message interfaces
arlab_template_interfaces	Template interfaces for custom message types
arlab_templates	ROS 2 package templates for new components

Perception & Vision

Package	Description	Documentation
arlab_computer_vision	YOLO-based object detection/segmentation integrated with the knowledge base	Read more
arlab_perception	Perception helpers (lidar relay, camera interface)	Read more

Knowledge & State Management

Package	Description	Documentation
arlab_knowledge	Database for handling robot surroundings, events and maps	Read more
arlab_knowledge_interfaces	Knowledge base ROS 2 message and service interfaces	read above

Decision Making & Behavior

Package	Description	Documentation
arlab_decision_making	Global decision maker using py_trees-based behavior trees	Read more
py_trees_error_selector	Error selector behavior tree node for robust error handling	read above

Manipulation

Package	Description	Documentation
arlab_manipulation	Orchestrates manipulation commands (move, pick, place, home, open, close)	Read more
arlab_manipulation_cpp	C++ interface for manipulation execution	read above
manipulator_description	URDF/xacro robot model, launch files, controller configuration	Read more
manipulator_robot_driver	Actual driver for UR robots (modified for ARLab)	Read more
manipulator_ur_moveit_config	Customized MoveIt 2 configuration for the manipulator	Read more

Movement & Navigation

Package	Description
arlab_movement	Movement behaviors and navigation

Speech & Audio

Package	Description	Documentation
arlab_speech_controller	Speech synthesis and audio processing (MOSHI TTS)	Read more
arlab_whisper	Speech-to-Text pipeline using faster_whisper	Read more

Safety

Package	Description
arlab_safety	Safety monitoring and emergency handling

Visualization & UI

Package	Description	Documentation
arlab_ui	Robot UI and visualization components	Read more
arlab_vizbox	Visualization box for robot state display (deprecated)

Contributing

Everybody can contribute. Get in touch with us for questions and open tasks. We are happy to have you on board. We follow the

For easy start we created a CONTRIBUTING Guideline.

Authors and acknowledgment

• Lukas Asam
• Daniel Gabler
• Aleksander Michalak
• Sofia Öttl
• Jonas Platzer
• Leonie Schmidt
• Peter Viechter
• Meruna Yugarajah
• Nils Mandischer
• Robert Jeutter