This is a 2D LiDAR SLAM course project designed for ROS1 Melodic + Python 2.7 + Slamtec Slamware ROS SDK.
The goal of this project is to build a complete 2D SLAM pipeline based on the LiDAR data provided by the Slamware ROS SDK.
Main functions include:
- Directly subscribe to the
/scantopic published byslamware_ros_sdk_server_node - Implement a scan-to-scan ICP front-end
- Add keyframes, loop closure detection, and pose graph optimization based on the ICP front-end
- Save the trajectory, loop closure edges, optimized results, and plotting results into the project-level
result/folder
Please read the following documents first:
上机步骤.md代码功能说明.mdREADME_上机步骤.mdREADME_代码说明.md离线依赖清单.txt
This project is intended to run on the Unitree Go1 robot with an Ubuntu-based ROS environment.
The main running environment is:
- Robot platform: Unitree Go1
- Operating system: Ubuntu with ROS1 Melodic
- ROS version: ROS1 Melodic
- Python version: Python 2.7
- LiDAR SDK: Slamtec Slamware ROS SDK
- LiDAR data source:
/scantopic fromslamware_ros_sdk_server_node
The SLAM algorithm, LiDAR topic subscription, ICP front-end, loop closure detection, and pose graph optimization are all executed inside the Go1 robot’s Ubuntu system.
During the experiment, I used a Windows laptop to remotely connect to the Unitree Go1 robot.
The general workflow is:
Windows laptop
↓ remote connection
Unitree Go1 Ubuntu system
↓ run ROS and Slamware SDK
slamware_ros_sdk_server_node publishes /scan
↓
this project subscribes to /scan and runs SLAM
Click to read
编代码时，我心尚平；
初试debug，亦可自持；一改、二改、三改，虽有疲惫，犹能修补。
我本以为，历经风霜，已学会成人之冷静克制。
岂料最后破我道心者，非算法，非代码，非环境依赖，竟是Go1的电池、雷达还有那该死的小米WiFi，靠。
在我修炼情绪稳定的路途上，Go1，你当真是功德无量。
往哪拜才能让你在上机的时候丝滑运行，人为什么要在没有真机的时候写上机代码，上机环境怎么能没有网，
gpt你看看你训练出来的兵，竟无人会古法编程哈哈哈哈哈哈，
还有你个小米wifi能不能不要老断了，那个薛定谔的雷达能不能别时好时坏了，
最朴素的解决方法竟然是拔USB我真是服了（绝望研究生belike）
2026/5/12补充：别看了姐妹，这版没运行成功，该死的python，还是玩c++去吧