低成本智能家居系统：树莓派/Orange Pi + MediaTek LinkIt + Azure云端

组件与物料清单

云平台服务

• Microsoft Azure 存储服务（可选，本地存储亦可）

系统架构概述

本项目演示如何搭建超低成本智能家居数据采集系统：

1. 传感器节点：各类传感器（温湿度、光照、气体、PIR、土壤湿度等）通过Arduino/MediaTek采集数据
2. 无线传输：XBee或廉价433MHz/LoRa模块将数据汇聚到MediaTek LinkIt网关
3. 边缘网关：MediaTek LinkIt封装JSON，通过WiFi POST至本地服务器
4. 本地预处理服务器：$15 Orange Pi运行Node.js，接收数据、添加UTC时间戳、签名、过滤清洗
5. 云端存储：本地服务器转发至Azure表存储（可选，也可仅本地存储）

数据模型：统一JSON格式

采用desired字段表示设备期望状态，便于前端同步：

单传感器示例

// 温湿度
{ "deviceid": "iot/myhome/livingroom/thermostat",
  "desired": { "temperature": "24", "humidity": "80" },
  "timeutc": "2015-11-23T00:41:48.824Z",
  "signedBy": "localserver" }

// 麦克风声强
{ "deviceid": "iot/myhome/livingroom/mic",
  "desired": { "intensity": "66" },
  "timeutc": "...", "signedBy": "localserver" }

// 超声波测距
{ "deviceid": "iot/myhome/livingroom/obstacle",
  "desired": { "proximity": "150" }, ... }

// PIR人体红外
{ "deviceid": "iot/myhome/livingroom/pir",
  "desired": { "motion_start": "2500", "motion_end": "5345" }, ... }

// GPS定位
{ "deviceid": "iot/mydevice/location",
  "desired": { "lat": "25.5", "long": "32.2" }, ... }

// 土壤湿度
{ "deviceid": "iot/myhome/garden/moisture",
  "desired": { "moisture": "20" }, ... }

// 燃气泄漏
{ "deviceid": "iot/myhome/kitchen/gas",
  "desired": { "leak": "No" }, ... }

// 光照强度
{ "deviceid": "iot/myhome/livingroom/light",
  "desired": { "intensity": "60" }, ... }

房间级聚合示例

// 客厅聚合
{ "deviceid": "/iot/myhome/livingroom",
  "desired": {
    "temperature": "25", "humidity": "80",
    "window": "open", "door": "closed",
    "light": "80", "noise": "low"
  }}

// 卧室聚合
{ "deviceid": "/iot/myhome/bedroom",
  "desired": {
    "temperature": "24", "humidity": "70",
    "window": "closed", "door": "closed",
    "light": "80", "ac": "on", "noise": "low"
  }}

MediaTek LinkIt 网关代码

完整代码见GitHub仓库，核心逻辑：连接WiFi → 循环读取/模拟传感器 → 构建JSON → HTTP POST到本地服务器 → 处理超时重连。

// 关键片段：发送温度数据
void updateTemperature() {
  DynamicJsonBuffer jsonBuffer;
  JsonObject& root = jsonBuffer.createObject();
  JsonObject& desired = root.createNestedObject("desired");
  desired["temperature"] = String(random(20,40));  // 替换为真实读数
  desired["humidity"] = String(random(60,80));
  root["deviceid"] = "/iot/myhome/livingroom/thermostat";
  root["timeutc"] = "";  // 由本地服务器填充
  root["signedby"] = "";
  
  String len = String(root.measureLength());
  c.println("POST /iot/myhome/livingroom/thermostat HTTP/1.1");
  c.println("Host: " SITE_URL);
  c.println("Content-Type: application/json");
  c.println("Cache-Control: no-cache");
  c.print("Content-Length: "); c.println(len);
  c.println("Connection: close"); c.println();
  root.printTo(c);
  // ... 等待响应超时处理
}

本地服务器

硬件：Orange Pi ($15) + Arch Linux ARM + Node.js

部署：

1. 克隆GitHub仓库 iotapp
2. npm install 安装依赖
3. 修改index.js中的云端API地址
4. 运行 node app.js 监听3000端口

Node.js 路由示例

router.post('/iot/myhome/livingroom/thermostat', function(req, res) {
  var state = req.body;
  state.timeutc = new Date();
  state.signedby = "orangepi";
  
  // 转发至Azure云端API
  request.post({
    url: 'https://your-azure-domain/api/iot/myhome/livingroom/thermostat',
    json: true, body: state
  }, function(err, resp, body) { ... });
  
  res.send('Living room temperature recorded');
});
// ... 其他传感器路由同理

云端ASP.NET Core Web API

Visual Studio 2015创建ASP.NET 5项目，Entity Framework写入Azure表存储。每类传感器对应一个Controller，示例：

[Route("api/iot/myhome/livingroom/[Controller]")]
public class ThermoStatController : IoTController {
  [HttpPost]
  public IActionResult Post([FromBody]IotTemperature data) {
    // 写入Azure Table Storage
    data.PartitionKey = "thermostat";
    data.RowKey = Guid.NewGuid().ToString();
    table.Execute(TableOperation.Insert(data));
    return Ok("Logged to Azure");
  }
}
// IotTemperature 含 deviceid, timeutc, signedby, desired{temperature,humidity}

完整Controller列表：ThermoStat, PIR, Moisture, LivingRoom, Light, GPS, Gas, Distance, BedRoom。

原理图与接线

项目源码仓库

GitHub: iot-mediatek-linkIt - 含MediaTek接收端、Node.js本地服务器、ASP.NET Core云端API全套代码

总结

• 证明了$15 SBC足以胜任IoT边缘预处理
• JSON统一数据模型便于扩展与前端绑定
• 本地服务器可离线运行，云端同步可选
• 代码有重构空间，欢迎PR改进