Team23_I2Ros_2025

Quick Project Summary

A quick project overview can be seen here:

https://youtu.be/YZ5LkFzoBeQ

Running the system

Installing Dependencies

In the team23_i2ros_2025 folder execute the following command:

./setup/setup_environment.sh

We decided against automatically rebooting your computer. Please reboot manually.

Quick Start

# Build the project
catkin build

# Start system
source devel/setup.bash
roslaunch system autonomous_system.launch

In the rare case of a black screen in rviz, please restart the simulation.

System Arguments

Argument	Default	Description
generate_octomap	false	Generate octomap from sensor data
enable_simulation	true	Enable simulation module
enable_decisionmaker	true	Enable decision maker
enable_perception	true	Enable perception module (camera transforms and depth processing)
enable_pose_logging	false	Enable pose logging
enable_path_posting	true	Enable path posting
enable_frenet_planner	false	Enable frenet optimal planner for trajectory planning
enable_dwa_planner	false	Enable DWA planner for trajectory planning
enable_teb_planner	true	Enable TEB planner for trajectory planning
enable_controller	true	Enable controller module
path_file	$(find pathplanner) /paths/octo_based/ smoothed_v1.csv	Path to path file to use as fallback in case the trajectory planner lacks compute
filter_distance	0.1	Minimum distance between path points (meters)
enable_rviz	true	Start RViz with default config
robot_model	$(find system) /urdf/ourcar.urdf	Path to robot URDF model
enable_model_display	true	Enable robot model and TF publishing
decisionmaker_output	log	Output type for decisionmaker nodes
perception_output	log	Output type for perception nodes
pathplanner_output	log	Output type for pathplanner nodes
controller_output	log	Output type for controller nodes
robot_state_publisher_output	log	Output type for robot_state_publisher
rviz_output	log	Output type for rviz

Examples

# Run simulation with controller
roslaunch system autonomous_system.launch enable_controller:=true

# Run without simulation (hardware mode)
roslaunch system autonomous_system.launch enable_simulation:=false

# Enable pose logging with custom fallback path
roslaunch system autonomous_system.launch enable_pose_logging:=true path_file:="$(rospack find system)/../../trajectories/custom_trajectory/my_trajectory.csv"

Project Structure

stateDiagram-v2
  state "Simulation" as Simulation {
    Perception: Perception
    DecisionMaking: Decision Making
    PathPlanner: Path Planner
    Control: Control
    PoseLogger: Pose Logger

    Perception --> PathPlanner: 2d occupancy map, <br> obstacle states, <br> pointcloud
    Perception --> DecisionMaking: 2d occupancy map, <br> obstacle states, <br> traffic light states
    PathPlanner --> Control: Trajectory
    DecisionMaking --> PathPlanner: brake command to stop at <br> traffic light and prevent <br> collisions
  }

  state "External Analysis" as Analysis {
    PoseAnalysis: min turning radius, <br> max speed/acceleration, <br> demo trajectory.. 
  }

  csvlogs: CSV Logs
  PoseLogger --> csvlogs

Loading

To preview Mermaid diagrams in VS Code, you may need the "Markdown Preview Mermaid Support" extension by Matt Bierner.

src/
├── system/         # System launch orchestration
├── simulation/     # Unity-based simulation
├── perception/     # 3D perception pipeline (point clouds, depth processing)
├── pathplanner/    # Path planning and trajectory handling
├── controller/     # Vehicle control for following trajectory
└── dummy_controller/ # Reference controller (not used in launch)

Further Inormation

Further Information about the single packages can be found in the README files of the according package source folders. For example: The information and explanaition of the perception package can be found in ../src/perception/README.md

Third-Party Code Integration

This repository includes code from external sources that has been integrated directly (not as git submodules) for easier team collaboration:

autoware_ai_messages

• Source: https://github.com/autowarefoundation/autoware_ai_messages.git
• Purpose: ROS message definitions for Autoware AI (ROS1 compatible)
• Modifications:
  • Removed git history for easier integration
  • Added CATKIN_IGNORE files to unused packages (only autoware_msgs is built)
• License: Apache 2.0 (see src/autoware_ai_messages/LICENSE)

frenet_optimal_planner

• Source: https://github.com/SS47816/frenet_optimal_planner.git
• Purpose: Frenet coordinate-based trajectory planner for autonomous vehicles
• Modifications:
  • Removed git history for easier integration
  • Integrated with Team23 system via custom adapter (src/pathplanner/scripts/planner_adapter.py)
• License: MIT (see src/frenet_optimal_planner/LICENSE)