g0plus_dockerfile/docker-assets/code/Hierarchical_System/README.md.zh

# Hierarchical System ROS2

## 0- 前言

### What we have

- 主要逻辑（python）文件夹及文件的路径及命名如下

```
src/
   └── g0_vlm_node/
        └── g0_vlm_node
            ├── utils/                  # 储存与Gemini api处理相关的func
            └── vlm_main.py             # VLM提供服务的核心逻辑
```
- 注：以上包内：
    - vlm_main.py

### 开发说明


- VLM
    1. 将String格式化，使得EHI发送的json字符串，改为结构化字符串
    2. 支持接收EHI的缓存开关
    3. 支持参数化启动，用`--use-qwen`和`--no-use-qwen`控制模型使用，默认是Gemini


## 1- Install

1. 安装Python依赖库

参考https://github.com/whitbrunn/G0


2. 编译工作空间

将`src/`文件夹clone到本地工作空间下`TO/YOUR/WORKSPACE/`，运行

```
cd TO/YOUR/WORKSPACE/
colcon build --symlink-install --cmake-args -DPython3_ROOT_DIR=$CONDA_PREFIX
```

Note:

用`ros2 pkg list | grep PACK_NAME` 检查是否有以下ROS包： 
- `g0_vlm_node`


## 2- Usage

1. 设置api key

```
export API_KEY=<YOUR_GEMINI_API_KEY> 
export API_KEY_QWEN=<YOUR_QWEN_API_KEY> 
```

2. 启动VLM Node

1.1 先按所在环境配置代理（Gemini之必需，若使用qwen版，请跳到1.3）

```
export https_proxy=http://127.0.0.1:<PORT>
export http_proxy=http://127.0.0.1:<PORT>
export all_proxy=http://127.0.0.1:<PORT>
```
1.2 验证外网是否可通

```
curl -I www.google.com
```

预期显示（部分），

```
HTTP/1.1 200 OK
Transfer-Encoding: chunked
Cache-Control: private
Connection: keep-alive
```

1.3 确定上一步OK后，启动VLM节点

```
ros2 run g0_vlm_node vlm_main
```

*若使用qwen模型推理
```
unset http_proxy
unset https_proxy
unset all_proxy
ros2 run g0_vlm_node vlm_main -- --use-qwen
```


## 3- What you expect

- VLM收到Send请求输出，e.g.,

```
2025-11-05 07:40:33.230 | INFO     | g0_vlm_node.vlm_main:vlm_processor1:153 - One hp successfully processed: 将咖啡罐用右手放到托盘上 -> [Low]: Pick up the coffee can with the right hand and place it on the tray.!
```

- VLM收到confirm请求，e.g.,

```
2025-11-05 07:40:47.641 | INFO     | g0_vlm_node.vlm_main:vlm_processor2:169 - One hp_ successfully sent to VLA: [Low]: Pick up the coffee can with the right hand and place it on the tray.!
```