STR Robot: Setup and Build Guide

Project Website
arXiv Paper

1. Prerequisites

• Ubuntu 22.04 LTS
• ROS 2 Humble Hawksbill
• Livox ROS Driver 2
• Micro-ROS Agent
• FAST-LIO / FAST-LIVO
• Nav2-based navigation stack

Install basic dependencies:

sudo apt update
sudo apt install -y \
    libpcl-dev \
    libeigen3-dev \
    python3-colcon-common-extensions \
    python3-rosdep \
    python3-vcstool \
    build-essential \
    cmake \
    git

Source ROS 2 Humble:

source /opt/ros/humble/setup.bash

---

2. USB Device Setup

Create udev rule for STM32:

sudo nano /etc/udev/rules.d/99-ttbot.rules

Paste the following content:

SUBSYSTEM=="tty", ATTRS{idVendor}=="1a86", ATTRS{idProduct}=="7523", SYMLINK+="ttbot_stm32", MODE="0666"

Apply udev rules:

sudo udevadm control --reload-rules && sudo udevadm trigger

Check device:

ls /dev/ttbot_stm32

---

3. Build Livox ROS Driver 2

cd ~/strbot_ws/src/livox_ros_driver2
source /opt/ros/humble/setup.bash
./build.sh humble
source install/setup.bash

---

4. Build STR Robot Workspace

cd ~/strbot_ws
source /opt/ros/humble/setup.bash
rosdep install --from-paths src --ignore-src -r -y
colcon build
source install/setup.bash

---

5. Micro-ROS Agent

ros2 run micro_ros_agent micro_ros_agent serial --dev /dev/ttbot_stm32

---

6. Simulation

Run each command in a separate terminal.

Terminal 1: Robot Bringup

cd ~/strbot_ws
source install/setup.bash
ros2 launch ttbot_bringup sim.launch.py

Terminal 2: FAST-LIO Simulation

cd ~/strbot_ws
source install/setup.bash
ros2 launch fast_lio fast_lio_sim.launch.py

Terminal 3: Navigation Simulation

cd ~/strbot_ws
source install/setup.bash
ros2 launch ttbot_navigation navigation_sim.launch.py

---

7. Real-World Deployment

Run the real robot system in the following order:

1. Launch Micro-ROS Agent.
2. Launch real-robot bringup.
3. Launch FAST-LIO / FAST-LIVO in real mode.
4. Launch the navigation stack with:

use_sim_time:=false

Before running navigation, check the map used by the Navigation map server.

---

Experimental Results

Physical experiments were conducted using a custom Ackermann-steered mobile robot.

Physical Robot

Real-World Path-Tracking Performance

The proposed architecture was deployed on the physical robot to verify sim-to-real consistency. A-GMPC consistently achieved lower RMSE than the standard MPC for all tested trajectories and speeds.

• At 0.6 m/s Average Speed: A-GMPC RMSE was 0.2349 m (Lemniscate) and 0.2588 m (Square); Standard MPC RMSE was 0.3194 m (Lemniscate) and 0.4260 m (Square).
• At 1.1 m/s Average Speed: A-GMPC RMSE was 0.3048 m (Lemniscate) and 0.2777 m (Square); Standard MPC RMSE was 0.3281 m (Lemniscate) and 0.4413 m (Square).
• At 1.5 m/s Average Speed: A-GMPC RMSE was 0.3582 m (Lemniscate) and 0.6182 m (Square); Standard MPC RMSE was 0.5149 m (Lemniscate) and 1.1475 m (Square).

Both controllers executed well within the 30 Hz (33.3 ms) control loop.

Runtime Statistics

Real-World Path Tracking

  

Video Demonstration: https://youtu.be/eXZOD7MUVX8.

Autonomous Navigation Demonstration

An end-to-end real-world navigation experiment validated the complete autonomy pipeline, successfully integrating onboard perception, mapping, path planning, and trajectory tracking. The system achieved smooth tracking behavior with a low tracking error of RMSE = 0.2624 m over a reference path length of 114.0290 m.

Environmental Mapping

  

Trajectory Tracking Performance

Video Demonstration: https://youtu.be/2TXuBDscRR4.

---

Sub-repositories

The core development stack of the STR Robot is modularized into the following specialized sub-repositories:

• Hardware (PCB Design): https://github.com/Nvinh5148/STR_PCB.git
• Firmware (micro-ROS Workspace): https://github.com/Nvinh5148/microros_ws.git