Autonomous Racing Robot (STM32 + ROS1)

STM32F446RE + ROS1 racing robot: RPLIDAR C1 + IMU, Bluetooth comms, hector_slam mapping, PID motor control, and a custom move_base local planner optimized for high-speed go-and-return navigation via referee node.

Author: [StarDust 星辰涵], Beijing University of Posts and Telecommunications (BUPT)
License: MIT (see LICENSE)

📑 目录 / Table of Contents

• Autonomous Racing Robot (STM32 + ROS1)
• 📄 项目说明 / Project Description
• 🎥 演示视频 / Demo Video
• 🛠 适用场景 / Use Cases
• 📋 设备清单 / Equipment List
• 🔧 消息通信架构 / Communication Architecture
• 📂 项目结构 / Project Structure
• ⚙️ 关键模块说明 / Key Modules Overview
• 🔁 移植建议 / Porting Guide
• 🐞 已知问题 / Known Issues

---

📄 项目说明 / Project Description

🇨🇳 中文说明

本项目是一个基于双轮差速驱动（辅以万向轮）的自动驾驶竞速小车，通过蓝牙实现 STM32 嵌入式系统与 ROS1（Noetic）上位机的紧密协同。嵌入式端实时采集 RPLIDAR C1 激光雷达与 IMU 数据，接收上位机下发的期望速度指令，执行高响应 PID 电机控制，并通过蓝牙进行双向通信。上位机端采用 hectorMapping_slam 实现无里程计的纯激光 SLAM 建图，并对 move_base 的局部规划器进行了定制化改造，摒弃传统避障逻辑，专为高速竞速路径跟踪优化。项目还包含一个“裁判节点”，用户只需输入相对于起点的目标坐标，小车即可自动导航至终点并返航。整体设计追求远程透传的slam建图、竞速和精准控制。

🇺🇸 English

This project presents an autonomous racing robot based on a differential-drive chassis (with a caster wheel), featuring tight integration between an STM32-based embedded system and a ROS1 (Noetic) navigation stack via Bluetooth. The embedded side handles real-time data acquisition from an RPLIDAR C1 and an IMU, executes PID-controlled motor commands based on velocity targets from the host, and communicates bidirectionally over Bluetooth. On the ROS side, hectorMapping_slam enables lidar-only SLAM for map building, while a customized local planner in move_base is optimized for high-speed racing trajectories—prioritizing path tracking over traditional obstacle avoidance. A dedicated “referee node” allows users to specify a goal coordinate relative to the start point, enabling fully autonomous go-and-return navigation. The overall design aims for transparent SLAM mapping, racing-oriented planning, and precise control.

---

🎥 演示视频 / Demo Video

🇨🇳 中文

• 完整演示（Bilibili）: 点击观看视频
• 快速预览-已经过倍速（GIF）:

🇺🇸 English

• Full Demo (Bilibili): Watch on Bilibili
• Quick Preview-times faster (GIF):

---

🛠 适用场景 / Use Cases

🇨🇳 中文

本项目遵循最小化与轻量化设计原则，代码结构精简，专注于在资源受限条件下完成特定任务。其典型适用场景包括：

• 未知静态迷宫中的自主探索与竞速：用户只需设定一个相对于起点的目标坐标，小车即可在无预载地图的情况下，边通过 hector_slam 实时建图，边规划路径前往终点并自动返航。整个过程无需人工干预，适用于封闭、静态但初始未知的赛道环境。
• 远程透传式 SLAM 架构参考：当激光雷达（如 RPLIDAR C1）部署在嵌入式端（STM32），而 SLAM 与导航计算需在远程 ROS1 主机完成时，本项目提供了一套基于蓝牙的低延迟、双向透传通信方案，可作为嵌入式传感器 + 远程计算平台协同系统的参考实现。
• 教学或竞赛原型开发：适合用于机器人学、嵌入式系统与 ROS 集成的教学演示，或作为“起点-目标-返航”类竞速任务的快速原型基础。

⚠️ 注意事项：
本项目未实现动态障碍物避障——为提升竞速性能，代价地图已移除代价的实时更新与滤波机制，因此仅适用于静态环境（无移动障碍物）。
若无需蓝牙透传、可直接通过串口连接雷达与主机，推荐使用官方 rplidar_ros 驱动，其稳定性与兼容性更佳。

🇺🇸 English

This project follows a minimalist and lightweight design philosophy, with streamlined code focused on accomplishing specific tasks under resource constraints. It is best suited for the following scenarios:

• Autonomous exploration and racing in unknown static mazes: Users only need to specify a goal coordinate relative to the starting point. The robot will then autonomously explore, build a map in real time using hector_slam, navigate to the goal, and return—without requiring a pre-loaded map. This makes it ideal for closed, static environments that are initially unknown.
• Reference implementation for remote SLAM via transparent transmission: When the LiDAR (e.g., RPLIDAR C1) is mounted on an embedded platform (STM32) while SLAM and navigation run on a remote ROS1 host, this project provides a low-latency, bidirectional Bluetooth-based transparent communication framework. It serves as a practical reference for embedded sensor + remote compute architectures.
• Educational or competition prototyping: Useful for teaching ROS-embedded integration, SLAM, and autonomous navigation, or as a rapid prototype for “go-to-goal-and-return” robotics challenges.

⚠️ Note:
This project does not support dynamic obstacle avoidance. To maximize racing performance, the costmap disables real-time cost updates and filtering. Therefore, it is only suitable for static environments (no moving obstacles).
If a direct serial connection between LiDAR and host is feasible (i.e., no Bluetooth relay needed), the official rplidar_ros driver is a more robust and maintainable choice.

---

📋 设备清单 / Equipment List

🇨🇳 中文

• NUCLEO-F446RE Nucleo-64 开发板
• MPU6050 6DOF 六轴 6 轴姿态加速度传感器模块(I2C)
• HC-04 初学者套餐
• RPLIDAR C1 思岚 C1 激光雷达 + 转接线 + 安装支架
• ATB236 双路驱动模块+12V 2550mAh 锂电池 3C 充电器套件【焊接排针】
• MCS20 带霍尔编码器(L:30 减速比) + 圆形三轮车底盘
• 杜邦线 20cm（40P/排/公对公）
• 电机 ph2.0 线连接 ATB236 驱动板 PH2.0 转 XH2.54
• 建图导航围栏道具（100 PCS）
• 收纳盒大号智能小车通用
• 平头螺丝 M2.5×8-头径 5
• 单头铜柱 M2.5×10+6 固定 8236 驱动板
• 单头铜柱 M2.5×50+5 固定雷达
• 平头 M3×12
• 六角螺帽 M3
• 双通铜柱 M3×30 固定开发板
• L 型固定 L30 固定开发板

🇺🇸 English

• NUCLEO-F446RE Nucleo-64 Development Board
• MPU6050 6DOF 6-axis Attitude and Acceleration Sensor Module (I²C)
• HC-04 Beginner Kit
• RPLIDAR C1 Slamtec LiDAR + Adapter Cable + Mounting Bracket
• ATB236 Dual Motor Driver Module + 12V 2550mAh Lithium Battery + 3C Charger Kit [Soldered Headers]
• MCS20 Motor with Hall Encoder (L:30 Gear Ratio) + Round 3-Wheel Robot Chassis
• 20cm Dupont Jumper Wires (40-pin/strip, male-to-male)
• Motor PH2.0 Cable Connector for ATB236 Driver (PH2.0 to XH2.54 adapter)
• Mapping & Navigation Barrier Props (100 pcs)
• Large Storage Box (Universal for Smart Robot Car)
• Flat-head Screw M2.5×8 (Head Diameter: 5mm)
• Single-end Brass Standoff M2.5×10+6 (for securing 8236 driver board)
• Single-end Brass Standoff M2.5×50+5 (for mounting LiDAR)
• Flat-head Screw M3×12
• Hex Nut M3
• Dual-end Brass Standoff M3×30 (for mounting development board)
• L-shaped Bracket L30 (for securing development board)

---

🔧 消息通信架构 / Communication Architecture

🇨🇳 中文

本项目采用极简透传通信模型，实现 STM32 嵌入式系统与 ROS1 主机之间的单向传感器上传与单向控制指令下发。整个系统无里程计、无闭环反馈，所有 SLAM 与导航计算均在 ROS 端基于纯激光数据完成。下图为系统消息流结构示意图。

图：系统整体消息流与模块交互关系

📡 通信流程说明：

1. 雷达数据透传（STM32 → ROS）：

   • RPLIDAR C1 输出的原始十六进制激光数据流，由 STM32 直接读取；
   • STM32 不解析雷达协议，仅在原始数据前添加自定义通信帧头（如起始标志、长度字段），通过蓝牙透明透传至 ROS 主机；
   • ROS 端的 /radar_parser_node 负责帧同步、去头、解析原始数据，并发布标准 /scan 消息供 hector_slam 使用。

2. 速度指令下发（ROS → STM32）：

   • 用户设定目标点后，定制版 move_base 输出期望的 /cmd_vel Twist 消息；
   • /velocity_parser_node 将轮速打包为轻量指令帧（左右轮速），通过蓝牙发送至 STM32；
   • STM32 接收后直接驱动电机执行 PID 控制，不回传任何状态或里程计信息。

3. 无闭环设计：

   • 嵌入式端完全屏蔽里程计数据的收发，不采集编码器、不计算位姿；
   • ROS 端使用 hector_slam 实现无里程计 SLAM（laser-only），不依赖 /odom；
   • 整个系统为开环控制架构，依赖高精度激光建图与路径跟踪，适用于静态、已知或可探索的迷宫环境。

✅ 该设计大幅降低嵌入式负载与通信开销，专注于远程透传 + 主机端 SLAM + 高速路径跟踪，契合轻量化竞速场景。

🇺🇸 English

This project adopts a minimalist transparent transmission model, enabling unidirectional sensor data upload and unidirectional control command download between the STM32 embedded system and the ROS1 host. The system operates without odometry and without closed-loop feedback—all SLAM and navigation are performed on the ROS side using laser-only data. The diagram below illustrates the complete message flow.

Figure: System-wide message flow and module interaction

📡 Communication Flow:

1. LiDAR Data Transparent Transmission (STM32 → ROS):

   • The raw hexadecimal data stream from RPLIDAR C1 is read directly by STM32;
   • STM32 does not parse the LiDAR protocol. Instead, it prepends a custom frame header (e.g., start flag, length field) and forwards the data transparently over Bluetooth to the ROS host;
   • On the ROS side, /radar_parser_node performs frame synchronization, header stripping, and protocol parsing, then publishes standard /scan messages for hector_slam.

2. Velocity Command Downlink (ROS → STM32):

   • After a goal is set, the customized move_base outputs desired /cmd_vel Twist messages;
   • /velocity_parser_node packs these velocities into lightweight command frames(left/right wheel velocities) and sends them via Bluetooth to STM32;
   • STM32 receives the commands and directly drives the motors using PID control, without sending back any status or odometry data.

3. Open-Loop Design:

   • The embedded side completely disables odometry transmission and reception—no encoder reading, no pose estimation;
   • ROS uses hector_slam for odometry-free SLAM (laser-only), with no reliance on /odom;
   • The entire system operates in an open-loop control mode, relying on accurate laser-based mapping and high-speed path tracking, making it suitable for static or explorable maze environments.

✅ This design significantly reduces embedded workload and communication overhead, focusing on remote transparent transmission + host-side SLAM + high-speed path tracking, ideal for lightweight racing scenarios.

---

📂 项目结构 / Project Structure

🇨🇳 中文

autonomous-car-project/
├── .gitignore
├── README.md
├── LICENSE
├── embedded/               # STM32CubeMX + Keil MDK 项目
│   ├── APP/                # 外设代码文件
│   ├── Core/               # 主程序main函数
│   ├── MDK-ARM/            # keil项目入口
│   └── MPU6050.ioc         # CubeMX 配置文件
├── ros_ws/
│   └── src/
│       ├── blue_teeth_pkg  # 蓝牙通信 + 雷达解析 + 控制中枢
│       ├── hector_nav_demo # SLAM + 导航
│       └── remoter_pkg     # 自定义键盘遥控
│       └── my_planner      # 自定义局部规划器 + 裁判节点
└── assets/                 # 图片、GIF 等资源

🇺🇸 English

autonomous-car-project/
├── .gitignore
├── README.md
├── LICENSE
├── embedded/               # STM32CubeMX + Keil MDK project
│   ├── APP/                # Peripheral driver code
│   ├── Core/               # Main application (main function)
│   ├── MDK-ARM/            # Keil project entry point
│   └── MPU6050.ioc         # CubeMX configuration file
├── ros_ws/
│   └── src/
│       ├── blue_teeth_pkg  # Bluetooth + radar parsing + control hub
│       ├── hector_nav_demo # SLAM + navigation
│       ├── remoter_pkg     # Custom keyboard teleoperation
│       └── my_planner      # Custom local planner + referee node
└── assets/                 # Images, GIFs

---

⚙️ 关键模块说明 / Key Modules Overview

🇨🇳 中文

嵌入式端（STM32F446RE）

ROS1 上位机

🇺🇸 English

Embedded Side (STM32F446RE)

ROS1 Host

---

🔁 移植建议 / Porting Guide

🇨🇳 中文

本项目分为 嵌入式下位机 与 ROS1 上位机 两部分，各自有独立的移植说明：

• 📌 嵌入式系统移植指南（中文）
• 📌 ROS1 上位机移植指南（中文）

请根据你的开发目标选择对应文档。

🇺🇸 English

This project consists of two parts: Embedded (lower-level) and ROS1 (upper-level). Each has its own porting guide:

• 📌 Embedded System Porting Guide (English)
• 📌 ROS1 Upper-Level Porting Guide (English)

Please refer to the relevant documentation based on your development target.

---

🐞 已知问题 / Known Issues

🇨🇳 中文

🇺🇸 English

---