{"id":110470,"date":"2026-09-28T18:24:50","date_gmt":"2026-09-28T10:24:50","guid":{"rendered":"https:\/\/www.wsisp.com\/helps\/110470.html"},"modified":"2026-09-28T18:24:50","modified_gmt":"2026-09-28T10:24:50","slug":"%e4%bb%8e-0-%e5%88%b0-1-%e6%9e%84%e5%bb%ba-iot-%e5%9c%ba%e6%99%af%e4%b8%8b%e7%9a%84-llm-agent%ef%bc%9a%e4%bb%bb%e5%8a%a1%e8%a7%84%e5%88%92%e3%80%81%e5%b7%a5%e5%85%b7%e8%b0%83%e7%94%a8%e4%b8%8e","status":"publish","type":"post","link":"https:\/\/www.wsisp.com\/helps\/110470.html","title":{"rendered":"\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218"},"content":{"rendered":"

1. \u5f15\u8a00<\/h3>\n

\u5f53\u524d LLM Agent \u5927\u591a\u805a\u7126\u7eaf\u8f6f\u4ef6\u573a\u666f，\u9002\u914d\u529e\u516c\u81ea\u52a8\u5316\u3001\u77e5\u8bc6\u5e93\u95ee\u7b54\u3001\u6570\u5b57\u4eba\u4ea4\u4e92\u7b49\u8f7b\u91cf\u5316\u9700\u6c42\u3002\u4e0d\u5c11\u5f00\u53d1\u8005\u76f4\u63a5\u5957\u7528\u901a\u7528 Agent \u67b6\u6784\u843d\u5730 IoT \u573a\u666f，\u6700\u7ec8\u9891\u53d1\u6307\u4ee4\u65e0\u6548\u3001\u8bbe\u5907\u51b2\u7a81\u3001\u72b6\u6001\u9519\u4e71\u3001\u91cd\u590d\u6267\u884c\u3001\u79bb\u7ebf\u5931\u6548\u7b49\u95ee\u9898，\u5bfc\u81f4\u9879\u76ee\u843d\u5730\u5361\u58f3\u3002<\/p>\n

IoT Agent \u7684\u6838\u5fc3\u7279\u8d28\u662f\u7269\u7406\u4e16\u754c\u95ed\u73af\u4ea4\u4e92，\u4e0e\u7eaf\u8f6f\u4ef6 Agent \u5b58\u5728\u672c\u8d28\u5dee\u5f02：\u7269\u7406\u8bbe\u5907\u5b58\u5728\u786c\u4ef6\u9650\u5236\u3001\u72b6\u6001\u5ef6\u8fdf\u3001\u64cd\u4f5c\u4e0d\u53ef\u9006\u3001\u7f51\u7edc\u6ce2\u52a8\u3001\u6307\u4ee4\u5e26\u5b9e\u4f53\u526f\u4f5c\u7528，\u65e0\u6cd5\u6cbf\u7528\u7eaf\u8f6f\u4ef6\u7684\u63a8\u7406\u4e0e\u6267\u884c\u903b\u8f91\u3002<\/p>\n

\u672c\u6587\u4ece\u96f6\u843d\u5730\u3001\u6452\u5f03\u7a7a\u6cdb\u7406\u8bba，\u624b\u628a\u624b\u642d\u5efa\u4e00\u5957\u53ef\u5546\u7528\u3001\u9ad8\u53ef\u9760\u7684 IoT LLM Agent \u67b6\u6784\u3002\u91cd\u70b9\u62c6\u89e3\u4e09\u5927\u6838\u5fc3\u80fd\u529b：IoT \u4e13\u5c5e\u4efb\u52a1\u89c4\u5212\u3001\u7269\u7406\u8bbe\u5907\u5de5\u5177\u8c03\u7528\u3001\u573a\u666f\u5316\u8bb0\u5fc6\u7ba1\u7406，\u642d\u914d\u53ef\u8fd0\u884c\u5b9e\u6218\u4ee3\u7801\u3001\u5b8c\u6574\u67b6\u6784\u8bbe\u8ba1\u4e0e\u843d\u5730\u907f\u5751\u65b9\u6848，\u5e2e\u52a9\u5f00\u53d1\u8005\u6253\u901a\u5927\u6a21\u578b\u4e0e\u7269\u8054\u7f51\u7684\u843d\u5730\u95ed\u73af\u3002<\/p>\n

\u6587\u7ae0\u6838\u5fc3\u5e72\u8d27：\u67b6\u6784\u62c6\u89e3\u3001\u6838\u5fc3\u539f\u7406\u3001\u53ef\u8fd0\u884c\u4ee3\u7801\u3001\u843d\u5730\u8e29\u5751\u3001\u4f18\u5316\u7b56\u7565\u3001\u751f\u4ea7\u7ea7\u90e8\u7f72\u601d\u8def，\u9002\u914d\u7269\u8054\u7f51\u5f00\u53d1\u3001\u5927\u6a21\u578b\u5e94\u7528\u5f00\u53d1\u3001\u8fb9\u7f18\u667a\u80fd\u843d\u5730\u5de5\u7a0b\u5e08\u5b66\u4e60\u4e0e\u590d\u7528\u3002<\/p>\n

2. \u4e3a\u4ec0\u4e48\u901a\u7528 LLM Agent \u65e0\u6cd5\u9002\u914d IoT \u573a\u666f？<\/h3>\n

\u843d\u5730\u5f00\u53d1\u524d，\u5fc5\u987b\u5398\u6e05\u901a\u7528 Agent \u4e0e IoT Agent \u7684\u6838\u5fc3\u5dee\u5f02，\u8fd9\u662f IoT Agent \u67b6\u6784\u8bbe\u8ba1\u7684\u5e95\u5c42\u903b\u8f91\u3002\u591a\u6570 IoT \u5927\u6a21\u578b\u9879\u76ee\u5931\u8d25，\u6839\u6e90\u90fd\u662f\u76f4\u63a5\u5957\u7528 OpenAI Function Call\u3001AutoGPT \u7b49\u901a\u7528\u8f6f\u4ef6 Agent \u67b6\u6784，\u5ffd\u89c6\u7269\u7406\u573a\u666f\u7ea6\u675f\u3002<\/p>\n

2.1 \u901a\u7528 Agent vs IoT Agent \u6838\u5fc3\u5dee\u5f02<\/h4>\n

\u901a\u7528 LLM Agent \u64cd\u4f5c\u5bf9\u8c61\u4e3a\u8f6f\u4ef6\u63a5\u53e3\u3001\u6587\u672c\u6570\u636e\u3001\u7f51\u7edc\u8d44\u6e90，\u652f\u6301\u91cd\u8bd5\u3001\u65e0\u5b9e\u4f53\u526f\u4f5c\u7528\u3001\u72b6\u6001\u53ef\u9006\u8f6c；\u800c IoT Agent \u9762\u5411\u4f20\u611f\u5668\u3001\u7ee7\u7535\u5668\u3001\u7535\u673a\u3001\u706f\u5149\u3001\u7a7a\u8c03\u7b49\u7269\u7406\u786c\u4ef6，\u573a\u666f\u7ea6\u675f\u66f4\u5f3a\u3001\u843d\u5730\u95e8\u69db\u66f4\u9ad8\u3002<\/p>\n\n\u5bf9\u6bd4\u7ef4\u5ea6\u901a\u7528\u8f6f\u4ef6 LLM AgentIoT \u573a\u666f LLM Agent<\/tr>\n\n\n\n\n\n\n\n
\u64cd\u4f5c\u5bf9\u8c61<\/td>\n\u8f6f\u4ef6 API\u3001\u6587\u672c\u3001\u6587\u4ef6\u3001\u6570\u636e\u5e93<\/td>\n\u7269\u7406\u8bbe\u5907\u3001\u4f20\u611f\u5668\u6570\u636e\u3001\u73af\u5883\u72b6\u6001<\/td>\n<\/tr>\n
\u64cd\u4f5c\u526f\u4f5c\u7528<\/td>\n\u65e0\u5b9e\u4f53\u526f\u4f5c\u7528，\u53ef\u91cd\u590d\u6267\u884c<\/td>\n\u7269\u7406\u72b6\u6001\u53d8\u66f4，\u4e0d\u53ef\u9006\u3001\u4e0d\u53ef\u968f\u610f\u91cd\u8bd5<\/td>\n<\/tr>\n
\u72b6\u6001\u7279\u6027<\/td>\n\u5b9e\u65f6\u540c\u6b65，\u65e0\u5ef6\u8fdf<\/td>\n\u8bbe\u5907\u72b6\u6001\u5ef6\u8fdf\u3001\u65ad\u8fde\u3001\u79bb\u7ebf\u3001\u72b6\u6001\u4e0d\u4e00\u81f4<\/td>\n<\/tr>\n
\u4efb\u52a1\u7ea6\u675f<\/td>\n\u903b\u8f91\u7ea6\u675f\u4e3a\u4e3b<\/td>\n\u786c\u4ef6\u7ea6\u675f\u3001\u65f6\u5e8f\u7ea6\u675f\u3001\u529f\u8017\u7ea6\u675f\u3001\u5b89\u5168\u7ea6\u675f<\/td>\n<\/tr>\n
\u6267\u884c\u4f18\u5148\u7ea7<\/td>\n\u65e0\u5f3a\u4f18\u5148\u7ea7<\/td>\n\u5b89\u5168\u4efb\u52a1\u4f18\u5148\u3001\u7d27\u6025\u4efb\u52a1\u63d2\u961f\u3001\u65e0\u6548\u4efb\u52a1\u8fc7\u6ee4<\/td>\n<\/tr>\n
\u5931\u8d25\u91cd\u8bd5<\/td>\n\u65e0\u9650\u91cd\u8bd5\u65e0\u98ce\u9669<\/td>\n\u91cd\u8bd5\u4f1a\u5bfc\u81f4\u8bbe\u5907\u8fc7\u8f7d\u3001\u6307\u4ee4\u51b2\u7a81，\u9700\u4e25\u683c\u9650\u5236<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

2.2 IoT LLM Agent \u7684\u6838\u5fc3\u8bbe\u8ba1\u76ee\u6807<\/h4>\n

\u7ed3\u5408\u4e0a\u8ff0\u5dee\u5f02，\u751f\u4ea7\u7ea7 IoT LLM Agent \u5fc5\u987b\u6ee1\u8db3\u56db\u5927\u6838\u5fc3\u8bbe\u8ba1\u76ee\u6807，\u5f7b\u5e95\u9002\u914d\u7269\u7406\u8bbe\u5907\u4ea4\u4e92\u573a\u666f：<\/p>\n

    \n
  • \u8bed\u4e49\u7406\u89e3\u7cbe\u51c6\u5316：\u5c06\u7528\u6237\u6a21\u7cca\u81ea\u7136\u8bed\u8a00（\u201c\u5c4b\u91cc\u6709\u70b9\u51b7，\u6536\u62fe\u4e00\u4e0b\u623f\u95f4\u201d）\u7cbe\u51c6\u62c6\u89e3\u4e3a\u53ef\u6267\u884c\u7684\u8bbe\u5907\u64cd\u4f5c\u4efb\u52a1<\/li>\n
  • \u4efb\u52a1\u6267\u884c\u5b89\u5168\u5316：\u89c4\u907f\u8bbe\u5907\u51b2\u7a81\u3001\u65e0\u6548\u6307\u4ee4\u3001\u5371\u9669\u64cd\u4f5c，\u675c\u7edd\u7269\u7406\u5b89\u5168\u98ce\u9669<\/li>\n
  • \u8bbe\u5907\u8c03\u7528\u6807\u51c6\u5316：\u7edf\u4e00\u9002\u914d\u706f\u5149\u3001\u7a7a\u8c03\u3001\u7a97\u5e18\u3001\u4f20\u611f\u5668\u7b49\u591a\u54c1\u7c7b IoT \u8bbe\u5907\u63a5\u53e3<\/li>\n
  • \u72b6\u6001\u8bb0\u5fc6\u6301\u4e45\u5316：\u8bb0\u4f4f\u8bbe\u5907\u5386\u53f2\u72b6\u6001\u3001\u7528\u6237\u4f7f\u7528\u4e60\u60ef\u3001\u573a\u666f\u6267\u884c\u8bb0\u5f55，\u5b9e\u73b0\u4e2a\u6027\u5316\u667a\u80fd\u4ea4\u4e92<\/li>\n<\/ul>\n

    3. \u4ece\u96f6\u642d\u5efa IoT LLM Agent \u6574\u4f53\u67b6\u6784<\/h3>\n

    \u672c\u6587\u6452\u5f03\u901a\u7528 Agent \u5197\u4f59\u8bbe\u8ba1，\u7ed3\u5408 IoT \u7269\u7406\u4ea4\u4e92\u7279\u6027，\u642d\u5efa\u4e00\u5957\u8f7b\u91cf\u5316\u3001\u9ad8\u53ef\u9760\u3001\u6613\u6269\u5c55\u7684\u4e94\u5c42\u843d\u5730\u67b6\u6784，\u4e5f\u662f\u5f53\u524d\u5bb6\u5c45\u3001\u5de5\u4e1a IoT \u9879\u76ee\u7684\u4e3b\u6d41\u5546\u7528\u67b6\u6784\u3002\u6574\u4f53\u6267\u884c\u94fe\u8def：\u7528\u6237\u8f93\u5165 \u2192 \u8bb0\u5fc6\u7ba1\u7406\u5c42 \u2192 \u4efb\u52a1\u89c4\u5212\u5c42 \u2192 \u5de5\u5177\u8c03\u7528\u5c42 \u2192 \u8bbe\u5907\u6267\u884c\u5c42 + \u72b6\u6001\u53cd\u9988\u5c42，\u5f62\u6210\u5b8c\u6574\u95ed\u73af\u3002<\/p>\n

    3.1 \u6574\u4f53\u67b6\u6784\u5206\u5c42\u8be6\u89e3<\/h4>\n

    \u7b2c\u4e00\u5c42：\u4ea4\u4e92\u8f93\u5165\u5c42<\/p>\n

    \u591a\u6a21\u6001\u6307\u4ee4\u5165\u53e3，\u652f\u6301\u7528\u6237\u6587\u5b57\u3001\u8bed\u97f3\u624b\u52a8\u8f93\u5165，\u540c\u65f6\u517c\u5bb9\u5b9a\u65f6\u89e6\u53d1\u3001\u4f20\u611f\u5668\u8054\u52a8\u81ea\u52a8\u89e6\u53d1（\u9ad8\u6e29\u542f\u7a7a\u8c03\u3001\u4eba\u4f53\u611f\u5e94\u5f00\u706f\u7b49），\u8986\u76d6 IoT \u5168\u573a\u666f\u6307\u4ee4\u6765\u6e90\u3002<\/p>\n

    \u7b2c\u4e8c\u5c42：\u8bb0\u5fc6\u7ba1\u7406\u5c42（IoT \u573a\u666f\u6838\u5fc3）<\/p>\n

    \u533a\u522b\u4e8e\u901a\u7528\u5bf9\u8bdd\u8bb0\u5fc6，IoT \u8bb0\u5fc6\u6838\u5fc3\u805a\u7126\u8bbe\u5907\u5b9e\u65f6\u72b6\u6001\u3001\u5386\u53f2\u64cd\u4f5c\u8bb0\u5f55\u3001\u7528\u6237\u4f7f\u7528\u504f\u597d\u3001\u56fa\u5b9a\u573a\u666f\u89c4\u5219，\u4e3a\u4efb\u52a1\u89c4\u5212\u63d0\u4f9b\u7cbe\u51c6\u72b6\u6001\u4f9d\u636e，\u89e3\u51b3\u8bbe\u5907\u72b6\u6001\u5ef6\u8fdf\u3001\u4e0a\u4e0b\u6587\u4e0d\u4e00\u81f4\u3001\u63a8\u7406\u5931\u771f\u7b49\u6838\u5fc3\u95ee\u9898\u3002<\/p>\n

    \u7b2c\u4e09\u5c42：\u4efb\u52a1\u89c4\u5212\u5c42（\u5927\u8111\u6838\u5fc3）<\/p>\n

    Agent \u6838\u5fc3\u51b3\u7b56\u6a21\u5757，\u4f9d\u6258\u5927\u6a21\u578b\u5b8c\u6210\u81ea\u7136\u8bed\u8a00\u89e3\u6790\u3001\u590d\u6742\u4efb\u52a1\u62c6\u89e3\u3001\u4f9d\u8d56\u5173\u7cfb\u5224\u65ad\u3001\u4f18\u5148\u7ea7\u6392\u5e8f，\u540c\u65f6\u5185\u7f6e\u65e0\u6548\u4efb\u52a1\u8fc7\u6ee4\u3001\u5b89\u5168\u5408\u89c4\u6821\u9a8c，\u5c06\u6a21\u7cca\u7528\u6237\u9700\u6c42\u8f6c\u5316\u4e3a\u53ef\u4e32\u884c\/\u5e76\u884c\u6267\u884c\u7684\u6807\u51c6\u5316\u8bbe\u5907\u6307\u4ee4\u3002<\/p>\n

    \u7b2c\u56db\u5c42：\u5de5\u5177\u8c03\u7528\u5c42（\u8bbe\u5907\u9002\u914d\u6838\u5fc3）<\/p>\n

    \u7edf\u4e00\u6807\u51c6\u5316\u8bbe\u5907\u5de5\u5177\u6c60，\u5c01\u88c5\u5168\u54c1\u7c7b IoT \u8bbe\u5907\u7684\u63a7\u5236\u3001\u67e5\u8be2 API，\u96c6\u6210\u53c2\u6570\u6821\u9a8c\u3001\u6743\u9650\u6821\u9a8c\u3001\u9632\u6296\u53bb\u91cd\u3001\u5f02\u5e38\u515c\u5e95\u673a\u5236，\u662f\u5927\u6a21\u578b\u5bf9\u63a5\u7269\u7406\u8bbe\u5907\u7684\u552f\u4e00\u5b89\u5168\u5165\u53e3\u3002<\/p>\n

    \u7b2c\u4e94\u5c42：\u8bbe\u5907\u6267\u884c\u4e0e\u53cd\u9988\u5c42<\/p>\n

    \u7531\u8fb9\u7f18\u7f51\u5173\u3001\u8bbe\u5907\u6a21\u7ec4\u627f\u63a5\u6307\u4ee4，\u5b8c\u6210\u7269\u7406\u52a8\u4f5c\u6267\u884c，\u5b9e\u65f6\u56de\u4f20\u8bbe\u5907\u8fd0\u884c\u72b6\u6001\u4e0e\u6267\u884c\u7ed3\u679c，\u540c\u6b65\u66f4\u65b0\u8bb0\u5fc6\u5c42\u7f13\u5b58，\u5b9e\u73b0"\u611f\u77e5-\u51b3\u7b56-\u6267\u884c-\u53cd\u9988"\u7684\u5b8c\u6574\u95ed\u73af\u3002<\/p>\n

    3.2 \u6838\u5fc3\u6280\u672f\u6808\u9009\u578b（\u751f\u4ea7\u7ea7）<\/h4>\n
      \n
    • \u5927\u6a21\u578b\u5e95\u5ea7：\u901a\u4e49\u5343\u95ee \/ LLaMA3（7B\/14B）\u8fb9\u7f18\u8f7b\u91cf\u5316\u6a21\u578b，\u517c\u987e\u63a8\u7406\u901f\u5ea6\u4e0e\u51c6\u786e\u7387，\u652f\u6301\u672c\u5730\u90e8\u7f72\u65e0\u7f51\u7edc\u4f9d\u8d56<\/li>\n
    • \u5f00\u53d1\u8bed\u8a00：Python 3.9+（\u751f\u6001\u5b8c\u5584，\u9002\u914d\u5927\u6a21\u578b\u63a8\u7406\u4e0e IoT \u63a5\u53e3\u5f00\u53d1）<\/li>\n
    • \u8bb0\u5fc6\u5b58\u50a8：Redis（\u5b9e\u65f6\u8bbe\u5907\u72b6\u6001\u7f13\u5b58）+ SQLite（\u957f\u671f\u5386\u53f2\u4efb\u52a1\u8bb0\u5fc6）<\/li>\n
    • \u8bbe\u5907\u901a\u4fe1：MQTT（\u4e3b\u6d41 IoT \u534f\u8bae）+ HTTP（\u5e73\u53f0\u63a5\u53e3）<\/li>\n
    • \u6846\u67b6：\u81ea\u7814\u8f7b\u91cf\u5316 Agent \u6838\u5fc3（\u6452\u5f03 LangChain \u5197\u4f59\u5c01\u88c5，\u964d\u4f4e\u8fb9\u7f18\u8bbe\u5907\u7b97\u529b\u6d88\u8017）<\/li>\n<\/ul>\n

      4. \u6838\u5fc3\u6a21\u5757\u4e00：IoT \u4e13\u5c5e\u8bb0\u5fc6\u7ba1\u7406\u7cfb\u7edf（\u843d\u5730\u5b9e\u6218）<\/h3>\n

      \u591a\u6570\u5f00\u53d1\u8005\u7684\u6838\u5fc3\u8bef\u533a：\u7167\u642c\u901a\u7528\u5bf9\u8bdd\u8bb0\u5fc6\u65b9\u6848\u505a IoT Agent\u3002\u5b9e\u5219 IoT \u8bb0\u5fc6\u7684\u6838\u5fc3\u4e0d\u662f\u7559\u5b58\u5bf9\u8bdd，\u800c\u662f\u7cbe\u51c6\u7559\u5b58\u8bbe\u5907\u72b6\u6001\u4e0e\u7528\u6237\u4e60\u60ef\u3002\u4e00\u65e6\u8bb0\u5fc6\u72b6\u6001\u6ede\u540e\u3001\u5931\u771f，\u5927\u6a21\u578b\u6240\u6709\u4efb\u52a1\u89c4\u5212\u90fd\u4f1a\u57fa\u4e8e\u865a\u5047\u6570\u636e\u63a8\u7406，\u76f4\u63a5\u5bfc\u81f4\u6307\u4ee4\u9519\u4e71\u3001\u6267\u884c\u5931\u6548\u3002<\/p>\n

      4.1 IoT \u8bb0\u5fc6\u4e09\u5927\u6838\u5fc3\u7c7b\u578b<\/h4>\n
    • \u77ac\u65f6\u72b6\u6001\u8bb0\u5fc6（\u77ed\u671f\u8bb0\u5fc6）：\u57fa\u4e8e Redis \u5b9e\u73b0\u7684\u77ed\u671f\u7f13\u5b58，\u5b9e\u65f6\u5b58\u50a8\u706f\u5149\u3001\u7a7a\u8c03\u3001\u7a97\u5e18\u3001\u4f20\u611f\u5668\u7684\u5728\u7ebf\u72b6\u6001\u4e0e\u8fd0\u884c\u53c2\u6570，\u9ed8\u8ba4 30 \u5206\u949f\u8fc7\u671f，\u89c4\u907f\u9891\u7e41\u67e5\u8be2\u8bbe\u5907\u5e26\u6765\u7684\u6027\u80fd\u635f\u8017，\u89e3\u51b3\u72b6\u6001\u5ef6\u8fdf\u95ee\u9898\u3002<\/li>\n
    • \u573a\u666f\u884c\u4e3a\u8bb0\u5fc6（\u4e2d\u671f\u8bb0\u5fc6）：\u6c89\u6dc0\u7528\u6237\u9ad8\u9891\u573a\u666f\u884c\u4e3a，\u5982\u56fa\u5b9a\u65f6\u6bb5\u5173\u706f\u3001\u7279\u5b9a\u6e29\u5ea6\u542f\u505c\u7a7a\u8c03\u3001\u5f52\u5bb6\u81ea\u52a8\u901a\u98ce\u7b49，\u4e3a Agent \u4e3b\u52a8\u667a\u80fd\u89e6\u53d1\u3001\u573a\u666f\u81ea\u52a8\u5316\u63d0\u4f9b\u6570\u636e\u652f\u6491\u3002<\/li>\n
    • \u5386\u53f2\u4efb\u52a1\u8bb0\u5fc6（\u957f\u671f\u8bb0\u5fc6）：\u57fa\u4e8e SQLite \u957f\u671f\u5b58\u50a8\u5386\u53f2\u6307\u4ee4\u3001\u6267\u884c\u7ed3\u679c\u3001\u5f02\u5e38\u62a5\u9519，\u7528\u4e8e\u95ee\u9898\u6eaf\u6e90\u3001\u884c\u4e3a\u5206\u6790\u4e0e\u6a21\u578b\u8fed\u4ee3\u4f18\u5316\u3002<\/li>\n

      4.2 \u8bb0\u5fc6\u7ba1\u7406\u6838\u5fc3\u4ee3\u7801\u5b9e\u73b0<\/h4>\n

      \u91c7\u7528 Redis + SQLite \u53cc\u5b58\u50a8\u67b6\u6784，\u517c\u987e\u72b6\u6001\u5b9e\u65f6\u6027\u4e0e\u6570\u636e\u6301\u4e45\u6027，\u8f7b\u91cf\u5316\u9002\u914d\u8fb9\u7f18\u8bbe\u5907\u4f4e\u7b97\u529b\u8fd0\u884c\u73af\u5883，\u5b8c\u6574\u53ef\u843d\u5730\u4ee3\u7801\u5982\u4e0b：<\/p>\n

      import<\/span> redis
      \nimport<\/span> sqlite3
      \nimport<\/span> time
      \nfrom<\/span> typing import<\/span> Dict,<\/span> Optional<\/p>\n

      # \u521d\u59cb\u5316 Redis（\u77ed\u671f\u72b6\u6001\u8bb0\u5fc6）<\/span>
      \nredis_client =<\/span> redis.<\/span>Redis(<\/span>host=<\/span>"127.0.0.1"<\/span>,<\/span> port=<\/span>6379<\/span>,<\/span> db=<\/span>0<\/span>,<\/span> decode_responses=<\/span>True<\/span>)<\/span><\/p>\n

      # \u521d\u59cb\u5316 SQLite（\u957f\u671f\u4efb\u52a1\u8bb0\u5fc6）<\/span>
      \nconn =<\/span> sqlite3.<\/span>connect(<\/span>"iot_agent_memory.db"<\/span>,<\/span> check_same_thread=<\/span>False<\/span>)<\/span>
      \ncursor =<\/span> conn.<\/span>cursor(<\/span>)<\/span><\/p>\n

      # \u521b\u5efa\u4efb\u52a1\u8bb0\u5fc6\u8868<\/span>
      \ncursor.<\/span>execute(<\/span>'''
      \nCREATE TABLE IF NOT EXISTS iot_task_memory (
      \n id INTEGER PRIMARY KEY AUTOINCREMENT,
      \n user_input TEXT,
      \n task_plan TEXT,
      \n tool_call_result TEXT,
      \n device_state TEXT,
      \n create_time TIMESTAMP
      \n)
      \n'''<\/span>)<\/span>
      \nconn.<\/span>commit(<\/span>)<\/span><\/p>\n

      class<\/span> IoTMemoryManager<\/span>:<\/span>
      \n """IoT Agent \u4e13\u5c5e\u8bb0\u5fc6\u7ba1\u7406\u5668"""<\/span>
      \n def<\/span> __init__<\/span>(<\/span>self)<\/span>:<\/span>
      \n self.<\/span>redis_expire =<\/span> 1800<\/span> # \u72b6\u6001\u7f13\u5b58 30 \u5206\u949f\u8fc7\u671f<\/span><\/p>\n

      # \u66f4\u65b0\u8bbe\u5907\u5b9e\u65f6\u72b6\u6001<\/span>
      \n def<\/span> update_device_state<\/span>(<\/span>self,<\/span> device_id:<\/span> str<\/span>,<\/span> state:<\/span> Dict)<\/span> –<\/span>><\/span> None<\/span>:<\/span>
      \n """state: \u8bbe\u5907\u72b6\u6001\u5b57\u5178 {'status': 'on\/off', 'value': 25}"""<\/span>
      \n redis_client.<\/span>hmset(<\/span>f"device:<\/span>{<\/span>device_id}<\/span><\/span>"<\/span><\/span>,<\/span> state)<\/span>
      \n redis_client.<\/span>expire(<\/span>f"device:<\/span>{<\/span>device_id}<\/span><\/span>"<\/span><\/span>,<\/span> self.<\/span>redis_expire)<\/span><\/p>\n

      # \u83b7\u53d6\u8bbe\u5907\u7f13\u5b58\u72b6\u6001<\/span>
      \n def<\/span> get_device_state<\/span>(<\/span>self,<\/span> device_id:<\/span> str<\/span>)<\/span> –<\/span>><\/span> Optional[<\/span>Dict]<\/span>:<\/span>
      \n return<\/span> redis_client.<\/span>hgetall(<\/span>f"device:<\/span>{<\/span>device_id}<\/span><\/span>"<\/span><\/span>)<\/span> or<\/span> None<\/span><\/p>\n

      # \u5199\u5165\u957f\u671f\u4efb\u52a1\u8bb0\u5fc6<\/span>
      \n def<\/span> save_task_memory<\/span>(<\/span>self,<\/span> user_input:<\/span> str<\/span>,<\/span> task_plan:<\/span> str<\/span>,<\/span> tool_result:<\/span> str<\/span>,<\/span> device_state:<\/span> str<\/span>)<\/span>:<\/span>
      \n cursor.<\/span>execute(<\/span>
      \n "INSERT INTO iot_task_memory (user_input, task_plan, tool_call_result, device_state, create_time) VALUES (?,?,?,?,?)"<\/span>,<\/span>
      \n (<\/span>user_input,<\/span> task_plan,<\/span> tool_result,<\/span> device_state,<\/span> int<\/span>(<\/span>time.<\/span>time(<\/span>)<\/span>)<\/span>)<\/span>
      \n )<\/span>
      \n conn.<\/span>commit(<\/span>)<\/span><\/p>\n

      # \u83b7\u53d6\u8fd1\u671f\u5386\u53f2\u4efb\u52a1（\u7528\u4e8e\u4e0a\u4e0b\u6587\u63a8\u7406）<\/span>
      \n def<\/span> get_recent_memory<\/span>(<\/span>self,<\/span> limit:<\/span> int<\/span> =<\/span> 5<\/span>)<\/span> –<\/span>><\/span> list<\/span>:<\/span>
      \n cursor.<\/span>execute(<\/span>"SELECT * FROM iot_task_memory ORDER BY create_time DESC LIMIT ?"<\/span>,<\/span> (<\/span>limit,<\/span>)<\/span>)<\/span>
      \n return<\/span> cursor.<\/span>fetchall(<\/span>)<\/span><\/p>\n

      4.3 \u8bb0\u5fc6\u6a21\u5757\u843d\u5730\u5173\u952e\u4f18\u5316\u70b9<\/h4>\n
        \n
      • \u72b6\u6001\u589e\u91cf\u66f4\u65b0：\u4ec5\u5728\u8bbe\u5907\u72b6\u6001\u53d8\u66f4\u65f6\u5237\u65b0\u7f13\u5b58，\u51cf\u5c11 IO \u8bfb\u5199\u538b\u529b，\u9002\u914d\u8fb9\u7f18\u4f4e\u7b97\u529b\u8bbe\u5907<\/li>\n
      • \u7f13\u5b58\u8fc7\u671f\u673a\u5236：\u77ed\u671f\u72b6\u6001\u5b9a\u65f6\u8fc7\u671f，\u5f7b\u5e95\u89c4\u907f\u810f\u6570\u636e\u5bfc\u81f4\u7684\u9519\u8bef\u6307\u4ee4<\/li>\n
      • \u4e0a\u4e0b\u6587\u8f7b\u91cf\u5316：\u63a8\u7406\u4ec5\u52a0\u8f7d\u8fd1\u671f\u6709\u6548\u4efb\u52a1\u8bb0\u5fc6，\u907f\u514d\u4e0a\u4e0b\u6587\u8fc7\u957f\u5f15\u53d1\u63a8\u7406\u5ef6\u8fdf\u3001\u7cbe\u5ea6\u4e0b\u964d\u95ee\u9898<\/li>\n<\/ul>\n

        5. \u6838\u5fc3\u6a21\u5757\u4e8c：IoT \u4e13\u5c5e\u4efb\u52a1\u89c4\u5212（\u89e3\u51b3\u5927\u6a21\u578b\u4e71\u6307\u4ee4\u95ee\u9898）<\/h3>\n

        \u4efb\u52a1\u89c4\u5212\u662f IoT Agent \u7684\u51b3\u7b56\u6838\u5fc3\u3002\u901a\u7528 Agent \u4ec5\u505a\u7eaf\u903b\u8f91\u4efb\u52a1\u62c6\u89e3，\u800c IoT \u573a\u666f\u5fc5\u987b\u53e0\u52a0\u786c\u4ef6\u7ea6\u675f\u3001\u65f6\u5e8f\u7ea6\u675f\u3001\u5b89\u5168\u7ea6\u675f\u3001\u8bbe\u5907\u4f9d\u8d56\u7ea6\u675f，\u4ece\u6839\u6e90\u675c\u7edd\u5927\u6a21\u578b\u4e0b\u53d1\u65e0\u6548\u3001\u51b2\u7a81\u3001\u5371\u9669\u7684\u7269\u7406\u8bbe\u5907\u6307\u4ee4\u3002<\/p>\n

        5.1 IoT \u4efb\u52a1\u89c4\u5212\u56db\u5927\u6838\u5fc3\u89c4\u5219<\/h4>\n
      • \u72b6\u6001\u524d\u7f6e\u6821\u9a8c\u89c4\u5219：\u6240\u6709\u4efb\u52a1\u62c6\u89e3\u524d，\u4f18\u5148\u8bfb\u53d6\u8bb0\u5fc6\u6a21\u5757\u7684\u6700\u65b0\u8bbe\u5907\u72b6\u6001，\u7981\u6b62\u57fa\u4e8e\u9ed8\u8ba4\u503c\u3001\u5386\u53f2\u65e7\u72b6\u6001\u63a8\u7406\u3002\u4f8b\u5982\u706f\u5149\u5df2\u5f00\u542f\u65f6，\u81ea\u52a8\u62e6\u622a\u91cd\u590d\u5f00\u706f\u6307\u4ee4\u3002<\/li>\n
      • \u4efb\u52a1\u4f9d\u8d56\u6392\u5e8f\u89c4\u5219：\u7269\u7406\u8bbe\u5907\u64cd\u4f5c\u5b58\u5728\u5f3a\u65f6\u5e8f\u4f9d\u8d56，\u5173\u8054\u4efb\u52a1\u5fc5\u987b\u4e32\u884c\u6267\u884c\u3002\u4f8b\u5982\u9700\u5148\u5f00\u7a97\u901a\u98ce，\u518d\u5f00\u542f\u7a7a\u8c03；\u5148\u5173\u95ed\u9ad8\u5371\u8bbe\u5907，\u518d\u542f\u52a8\u7167\u660e\u8bbe\u5907\u3002<\/li>\n
      • \u5b89\u5168\u8fc7\u6ee4\u89c4\u5219：\u5185\u7f6e\u8bbe\u5907\u64cd\u4f5c\u9ed1\u540d\u5355，\u81ea\u52a8\u62e6\u622a\u9ad8\u6e29\u5f3a\u5236\u5236\u70ed\u3001\u53c2\u6570\u8d85\u9650\u8c03\u8282\u3001\u591c\u95f4\u9ad8\u5371\u8bbe\u5907\u542f\u52a8\u7b49\u98ce\u9669\u64cd\u4f5c，\u4fdd\u969c\u8bbe\u5907\u4e0e\u573a\u666f\u5b89\u5168\u3002<\/li>\n
      • \u4efb\u52a1\u5408\u5e76\u53bb\u91cd\u89c4\u5219：\u77ed\u65f6\u95f4\u5185\u540c\u4e00\u8bbe\u5907\u7684\u591a\u6761\u540c\u7c7b\u6307\u4ee4\u81ea\u52a8\u5408\u5e76\u53bb\u91cd，\u4f8b\u5982\u5c06"\u5f00\u706f\u3001\u8c03\u4eae\u706f\u5149"\u6574\u5408\u4e3a\u5355\u6761\u6307\u4ee4，\u51cf\u5c11\u7f51\u5173\u901a\u4fe1\u9891\u6b21\u4e0e\u8bbe\u5907\u8d1f\u8f7d\u3002<\/li>\n

        5.2 \u4efb\u52a1\u89c4\u5212 Prompt \u5de5\u7a0b（\u751f\u4ea7\u7ea7\u53ef\u7528）<\/h4>\n

        \u9488\u5bf9 IoT \u7269\u7406\u573a\u666f\u5b9a\u5236\u4e13\u5c5e\u7cfb\u7edf\u63d0\u793a\u8bcd，\u5f3a\u7ea6\u675f\u5927\u6a21\u578b\u63a8\u7406\u903b\u8f91，\u65e0\u9700\u6a21\u578b\u5fae\u8c03\u5373\u53ef\u5927\u5e45\u63d0\u5347\u4efb\u52a1\u89c4\u5212\u51c6\u786e\u7387，\u751f\u4ea7\u7ea7\u53ef\u7528 Prompt \u5982\u4e0b：<\/p>\n

        IOT_AGENT_SYSTEM_PROMPT =<\/span> """
        \n\u4f60\u662f\u4e13\u4e1a\u7684 IoT \u8bbe\u5907\u667a\u80fd\u8c03\u5ea6 Agent，\u4ec5\u8d1f\u8d23\u89e3\u6790\u7528\u6237\u6307\u4ee4\u5e76\u751f\u6210\u5408\u89c4\u7684\u8bbe\u5907\u6267\u884c\u4efb\u52a1，\u4e25\u683c\u9075\u5b88\u4ee5\u4e0b\u89c4\u5219：
        \n1. \u5fc5\u987b\u4f18\u5148\u8bfb\u53d6\u5f53\u524d\u8bbe\u5907\u5b9e\u65f6\u72b6\u6001，\u7981\u6b62\u91cd\u590d\u6267\u884c\u65e0\u6548\u64cd\u4f5c（\u8bbe\u5907\u5df2\u5f00\u542f\u5219\u4e0d\u518d\u4e0b\u53d1\u5f00\u542f\u6307\u4ee4）
        \n2. \u6240\u6709\u4efb\u52a1\u5fc5\u987b\u9002\u914d\u7269\u7406\u8bbe\u5907\u7ea6\u675f，\u53c2\u6570\u5728\u8bbe\u5907\u5408\u6cd5\u8303\u56f4\u5185：\u7a7a\u8c03\u6e29\u5ea6 16-30\u2103，\u706f\u5149\u4eae\u5ea6 0-100%
        \n3. \u4e25\u683c\u89c4\u907f\u5371\u9669\u64cd\u4f5c：\u7981\u6b62\u9ad8\u6e29\u5236\u70ed\u3001\u7981\u6b62\u8bbe\u5907\u53c2\u6570\u8d85\u9650\u3001\u7981\u6b62\u9ad8\u5371\u8bbe\u5907\u591c\u95f4\u8fdd\u89c4\u542f\u52a8
        \n4. \u5b58\u5728\u65f6\u5e8f\u4f9d\u8d56\u7684\u4efb\u52a1\u5fc5\u987b\u4e32\u884c\u6267\u884c，\u65e0\u4f9d\u8d56\u4efb\u52a1\u53ef\u5e76\u884c\u6267\u884c
        \n5. \u6700\u7ec8\u8f93\u51fa\u4ec5\u4e3a\u6807\u51c6\u5316\u4efb\u52a1\u5217\u8868，\u65e0\u591a\u4f59\u89e3\u91ca\u3001\u65e0\u53e3\u8bed\u5316\u5185\u5bb9，\u683c\u5f0f\u4e3a JSON \u6570\u7ec4
        \n6. \u81ea\u52a8\u5408\u5e76\u91cd\u590d\u4efb\u52a1\u3001\u8fc7\u6ee4\u65e0\u6548\u4efb\u52a1\u3001\u62e6\u622a\u98ce\u9669\u4efb\u52a1
        \n"""<\/span><\/p>\n

        5.3 \u4efb\u52a1\u89c4\u5212\u6838\u5fc3\u4ee3\u7801<\/h4>\n

        import<\/span> json
        \nfrom<\/span> typing import<\/span> List,<\/span> Dict
        \nfrom<\/span> llm_client import<\/span> local_llm_infer # \u672c\u5730\u5927\u6a21\u578b\u63a8\u7406\u63a5\u53e3<\/span>
        \nfrom<\/span> memory_manager import<\/span> IoTMemoryManager<\/p>\n

        memory =<\/span> IoTMemoryManager(<\/span>)<\/span><\/p>\n

        class<\/span> IoTTaskPlanner<\/span>:<\/span>
        \n """IoT \u4efb\u52a1\u89c4\u5212\u5668：\u9700\u6c42\u89e3\u6790\u3001\u4efb\u52a1\u62c6\u89e3\u3001\u5b89\u5168\u8fc7\u6ee4\u3001\u6392\u5e8f"""<\/span>
        \n def<\/span> __init__<\/span>(<\/span>self)<\/span>:<\/span>
        \n self.<\/span>system_prompt =<\/span> IOT_AGENT_SYSTEM_PROMPT<\/p>\n

        def<\/span> build_prompt<\/span>(<\/span>self,<\/span> user_input:<\/span> str<\/span>)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
        \n # \u62fc\u63a5\u8bbe\u5907\u72b6\u6001 + \u5386\u53f2\u8bb0\u5fc6 + \u7528\u6237\u6307\u4ee4<\/span>
        \n device_states =<\/span> {<\/span>dev:<\/span> memory.<\/span>get_device_state(<\/span>dev)<\/span> for<\/span> dev in<\/span> [<\/span>"light_01"<\/span>,<\/span> "air_01"<\/span>,<\/span> "curtain_01"<\/span>]<\/span>}<\/span>
        \n recent_memory =<\/span> memory.<\/span>get_recent_memory(<\/span>3<\/span>)<\/span>
        \n prompt =<\/span> f"""
        \n\u3010\u5f53\u524d\u8bbe\u5907\u72b6\u6001\u3011：<\/span>{<\/span>json.<\/span>dumps(<\/span>device_states,<\/span> ensure_ascii=<\/span>False<\/span>)<\/span>}<\/span><\/span>
        \n\u3010\u8fd1\u671f\u6267\u884c\u8bb0\u5f55\u3011：<\/span>{<\/span>recent_memory}<\/span><\/span>
        \n\u3010\u7528\u6237\u6307\u4ee4\u3011：<\/span>{<\/span>user_input}<\/span><\/span>
        \n\u8bf7\u6309\u7167\u89c4\u5219\u751f\u6210\u6807\u51c6\u5316\u8bbe\u5907\u4efb\u52a1 JSON \u6570\u7ec4\u3002
        \n"""<\/span><\/span>
        \n return<\/span> prompt<\/p>\n

        def<\/span> plan_task<\/span>(<\/span>self,<\/span> user_input:<\/span> str<\/span>)<\/span> –<\/span>><\/span> List[<\/span>Dict]<\/span>:<\/span>
        \n # 1. \u6784\u5efa\u63a8\u7406 prompt<\/span>
        \n prompt =<\/span> self.<\/span>build_prompt(<\/span>user_input)<\/span>
        \n # 2. \u5927\u6a21\u578b\u63a8\u7406\u4efb\u52a1\u89c4\u5212<\/span>
        \n llm_res =<\/span> local_llm_infer(<\/span>self.<\/span>system_prompt,<\/span> prompt)<\/span>
        \n # 3. \u89e3\u6790\u4efb\u52a1\u5e76\u57fa\u7840\u6821\u9a8c<\/span>
        \n try<\/span>:<\/span>
        \n task_list =<\/span> json.<\/span>loads(<\/span>llm_res)<\/span>
        \n except<\/span>:<\/span>
        \n return<\/span> [<\/span>]<\/span>
        \n # 4. \u4efb\u52a1\u5b89\u5168\u8fc7\u6ee4\u4e0e\u53bb\u91cd<\/span>
        \n valid_tasks =<\/span> self.<\/span>filter_illegal_tasks(<\/span>task_list)<\/span>
        \n return<\/span> valid_tasks<\/p>\n

        def<\/span> filter_illegal_tasks<\/span>(<\/span>self,<\/span> tasks:<\/span> List[<\/span>Dict]<\/span>)<\/span> –<\/span>><\/span> List[<\/span>Dict]<\/span>:<\/span>
        \n """\u8fc7\u6ee4\u975e\u6cd5\u3001\u65e0\u6548\u3001\u51b2\u7a81\u4efb\u52a1"""<\/span>
        \n valid =<\/span> [<\/span>]<\/span>
        \n for<\/span> task in<\/span> tasks:<\/span>
        \n device_id =<\/span> task.<\/span>get(<\/span>"device_id"<\/span>)<\/span>
        \n action =<\/span> task.<\/span>get(<\/span>"action"<\/span>)<\/span>
        \n params =<\/span> task.<\/span>get(<\/span>"params"<\/span>,<\/span> {<\/span>}<\/span>)<\/span>
        \n # \u72b6\u6001\u6821\u9a8c：\u8fc7\u6ee4\u65e0\u6548\u64cd\u4f5c<\/span>
        \n state =<\/span> memory.<\/span>get_device_state(<\/span>device_id)<\/span>
        \n if<\/span> not<\/span> state:<\/span>
        \n continue<\/span>
        \n if<\/span> state.<\/span>get(<\/span>"status"<\/span>)<\/span> ==<\/span> "on"<\/span> and<\/span> action ==<\/span> "open"<\/span>:<\/span>
        \n continue<\/span>
        \n if<\/span> state.<\/span>get(<\/span>"status"<\/span>)<\/span> ==<\/span> "off"<\/span> and<\/span> action ==<\/span> "close"<\/span>:<\/span>
        \n continue<\/span>
        \n # \u53c2\u6570\u5b89\u5168\u6821\u9a8c<\/span>
        \n if<\/span> device_id ==<\/span> "air_01"<\/span> and<\/span> params.<\/span>get(<\/span>"temp"<\/span>)<\/span>:<\/span>
        \n if<\/span> not<\/span> (<\/span>16<\/span> <=<\/span> params[<\/span>"temp"<\/span>]<\/span> <=<\/span> 30<\/span>)<\/span>:<\/span>
        \n continue<\/span>
        \n valid.<\/span>append(<\/span>task)<\/span>
        \n return<\/span> valid<\/p>\n

        6. \u6838\u5fc3\u6a21\u5757\u4e09：IoT \u5de5\u5177\u8c03\u7528\u4f53\u7cfb（\u6807\u51c6\u5316\u8bbe\u5907\u8c03\u5ea6）<\/h3>\n

        \u5de5\u5177\u8c03\u7528\u662f Agent \u843d\u5730\u7269\u7406\u8bbe\u5907\u7684\u552f\u4e00\u6267\u884c\u51fa\u53e3\u3002\u901a\u7528 Function Call \u4ec5\u4fdd\u8bc1\u63a5\u53e3\u53ef\u8c03\u7528，\u800c IoT \u5de5\u5177\u8c03\u7528\u9700\u540c\u65f6\u6ee1\u8db3\u8c03\u7528\u5408\u89c4\u3001\u6267\u884c\u53ef\u9760\u3001\u5f02\u5e38\u53ef\u515c\u5e95\u3001\u72b6\u6001\u53ef\u540c\u6b65\u56db\u5927\u751f\u4ea7\u7ea7\u8981\u6c42\u3002<\/p>\n

        6.1 IoT \u5de5\u5177\u6c60\u8bbe\u8ba1\u7406\u5ff5<\/h4>\n

        \u6452\u5f03\u901a\u7528 Agent \u7684\u52a8\u6001\u968f\u610f\u8c03\u7528\u6a21\u5f0f，\u91c7\u7528\u9759\u6001\u6807\u51c6\u5316\u5de5\u5177\u6c60：\u63d0\u524d\u5c01\u88c5\u6240\u6709\u8bbe\u5907\u7684\u53ef\u6267\u884c\u64cd\u4f5c\u3001\u53c2\u6570\u9608\u503c\u4e0e\u5f02\u5e38\u515c\u5e95\u7b56\u7565，\u5927\u6a21\u578b\u4ec5\u80fd\u8c03\u7528\u9884\u5b9a\u4e49\u5de5\u5177，\u5f7b\u5e95\u675c\u7edd\u975e\u6cd5\u8c03\u7528\u3001\u53c2\u6570\u5f02\u5e38\u95ee\u9898\u3002<\/p>\n

        \u6807\u51c6\u5316\u5de5\u5177\u7edf\u4e00\u7ed3\u6784：\u8bbe\u5907 ID + \u64cd\u4f5c\u7c7b\u578b + \u7ea6\u675f\u53c2\u6570 + \u8d85\u65f6\u673a\u5236 + \u91cd\u8bd5\u7b56\u7565，\u5b9e\u73b0\u8bbe\u5907\u8c03\u7528\u89c4\u8303\u5316\u3001\u7edf\u4e00\u5316\u3002<\/p>\n

        6.2 \u5de5\u5177\u8c03\u7528\u6838\u5fc3\u4ee3\u7801（MQTT \u8bbe\u5907\u901a\u4fe1）<\/h4>\n

        import<\/span> json
        \nimport<\/span> time
        \nfrom<\/span> typing import<\/span> Dict
        \nimport<\/span> paho.<\/span>mqtt.<\/span>client as<\/span> mqtt<\/p>\n

        # MQTT \u8bbe\u5907\u7f51\u5173\u914d\u7f6e<\/span>
        \nMQTT_BROKER =<\/span> "127.0.0.1"<\/span>
        \nMQTT_PORT =<\/span> 1883<\/span>
        \nMQTT_TOPIC_PREFIX =<\/span> "iot\/device\/"<\/span><\/p>\n

        # \u521d\u59cb\u5316 MQTT \u5ba2\u6237\u7aef<\/span>
        \nmqtt_client =<\/span> mqtt.<\/span>Client(<\/span>)<\/span>
        \nmqtt_client.<\/span>connect(<\/span>MQTT_BROKER,<\/span> MQTT_PORT,<\/span> 60<\/span>)<\/span><\/p>\n

        class<\/span> IoTToolCaller<\/span>:<\/span>
        \n """IoT \u8bbe\u5907\u5de5\u5177\u8c03\u7528\u5668"""<\/span>
        \n def<\/span> __init__<\/span>(<\/span>self)<\/span>:<\/span>
        \n # \u9884\u5b9a\u4e49\u6807\u51c6\u5316\u5de5\u5177\u6c60<\/span>
        \n self.<\/span>tool_pool =<\/span> {<\/span>
        \n "light_control"<\/span>:<\/span> self.<\/span>light_control,<\/span>
        \n "air_control"<\/span>:<\/span> self.<\/span>air_control,<\/span>
        \n "curtain_control"<\/span>:<\/span> self.<\/span>curtain_control
        \n }<\/span><\/p>\n

        def<\/span> light_control<\/span>(<\/span>self,<\/span> params:<\/span> Dict)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
        \n """\u706f\u5149\u63a7\u5236：\u5f00\u5173\u3001\u4eae\u5ea6\u8c03\u8282"""<\/span>
        \n device_id =<\/span> params[<\/span>"device_id"<\/span>]<\/span>
        \n payload =<\/span> json.<\/span>dumps(<\/span>{<\/span>
        \n "action"<\/span>:<\/span> params[<\/span>"action"<\/span>]<\/span>,<\/span>
        \n "brightness"<\/span>:<\/span> params.<\/span>get(<\/span>"brightness"<\/span>,<\/span> 100<\/span>)<\/span>,<\/span>
        \n "timestamp"<\/span>:<\/span> int<\/span>(<\/span>time.<\/span>time(<\/span>)<\/span>)<\/span>
        \n }<\/span>)<\/span>
        \n mqtt_client.<\/span>publish(<\/span>f"<\/span>{<\/span>MQTT_TOPIC_PREFIX}<\/span><\/span>{<\/span>device_id}<\/span><\/span>\/cmd"<\/span><\/span>,<\/span> payload)<\/span>
        \n return<\/span> f"\u706f\u5149<\/span>{<\/span>device_id}<\/span><\/span>\u64cd\u4f5c\u6267\u884c\u6210\u529f"<\/span><\/span><\/p>\n

        def<\/span> air_control<\/span>(<\/span>self,<\/span> params:<\/span> Dict)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
        \n """\u7a7a\u8c03\u63a7\u5236：\u5f00\u5173\u3001\u6e29\u5ea6\u8c03\u8282\u3001\u6a21\u5f0f\u5207\u6362"""<\/span>
        \n device_id =<\/span> params[<\/span>"device_id"<\/span>]<\/span>
        \n payload =<\/span> json.<\/span>dumps(<\/span>{<\/span>
        \n "action"<\/span>:<\/span> params[<\/span>"action"<\/span>]<\/span>,<\/span>
        \n "temp"<\/span>:<\/span> params.<\/span>get(<\/span>"temp"<\/span>,<\/span> 26<\/span>)<\/span>,<\/span>
        \n "mode"<\/span>:<\/span> params.<\/span>get(<\/span>"mode"<\/span>,<\/span> "auto"<\/span>)<\/span>,<\/span>
        \n "timestamp"<\/span>:<\/span> int<\/span>(<\/span>time.<\/span>time(<\/span>)<\/span>)<\/span>
        \n }<\/span>)<\/span>
        \n mqtt_client.<\/span>publish(<\/span>f"<\/span>{<\/span>MQTT_TOPIC_PREFIX}<\/span><\/span>{<\/span>device_id}<\/span><\/span>\/cmd"<\/span><\/span>,<\/span> payload)<\/span>
        \n return<\/span> f"\u7a7a\u8c03<\/span>{<\/span>device_id}<\/span><\/span>\u64cd\u4f5c\u6267\u884c\u6210\u529f"<\/span><\/span><\/p>\n

        def<\/span> curtain_control<\/span>(<\/span>self,<\/span> params:<\/span> Dict)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
        \n """\u7a97\u5e18\u63a7\u5236：\u5f00\u5408\u8c03\u8282"""<\/span>
        \n device_id =<\/span> params[<\/span>"device_id"<\/span>]<\/span>
        \n payload =<\/span> json.<\/span>dumps(<\/span>{<\/span>
        \n "action"<\/span>:<\/span> params[<\/span>"action"<\/span>]<\/span>,<\/span>
        \n "percent"<\/span>:<\/span> params.<\/span>get(<\/span>"percent"<\/span>,<\/span> 100<\/span>)<\/span>,<\/span>
        \n "timestamp"<\/span>:<\/span> int<\/span>(<\/span>time.<\/span>time(<\/span>)<\/span>)<\/span>
        \n }<\/span>)<\/span>
        \n mqtt_client.<\/span>publish(<\/span>f"<\/span>{<\/span>MQTT_TOPIC_PREFIX}<\/span><\/span>{<\/span>device_id}<\/span><\/span>\/cmd"<\/span><\/span>,<\/span> payload)<\/span>
        \n return<\/span> f"\u7a97\u5e18<\/span>{<\/span>device_id}<\/span><\/span>\u64cd\u4f5c\u6267\u884c\u6210\u529f"<\/span><\/span><\/p>\n

        def<\/span> execute_tool<\/span>(<\/span>self,<\/span> task:<\/span> Dict)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
        \n """\u7edf\u4e00\u5de5\u5177\u6267\u884c\u5165\u53e3"""<\/span>
        \n tool_name =<\/span> task[<\/span>"tool_name"<\/span>]<\/span>
        \n params =<\/span> task[<\/span>"params"<\/span>]<\/span>
        \n if<\/span> tool_name not<\/span> in<\/span> self.<\/span>tool_pool:<\/span>
        \n return<\/span> f"\u5de5\u5177<\/span>{<\/span>tool_name}<\/span><\/span>\u4e0d\u5b58\u5728，\u6267\u884c\u5931\u8d25"<\/span><\/span>
        \n # \u8d85\u65f6\u515c\u5e95，\u9632\u6b62\u963b\u585e<\/span>
        \n try<\/span>:<\/span>
        \n res =<\/span> self.<\/span>tool_pool[<\/span>tool_name]<\/span>(<\/span>params)<\/span>
        \n return<\/span> res
        \n except<\/span> Exception as<\/span> e:<\/span>
        \n return<\/span> f"\u8bbe\u5907\u8c03\u7528\u5f02\u5e38：<\/span>{<\/span>str<\/span>(<\/span>e)<\/span>}<\/span><\/span>"<\/span><\/span><\/p>\n

        6.3 \u5de5\u5177\u8c03\u7528\u751f\u4ea7\u7ea7\u4f18\u5316\u7b56\u7565<\/h4>\n
          \n
        • \u9632\u6296\u9650\u6d41：\u5355\u8bbe\u5907 1 \u79d2\u5185\u4ec5\u54cd\u5e94\u4e00\u6b21\u6307\u4ee4，\u907f\u514d\u9ad8\u9891\u8c03\u5ea6\u5bfc\u81f4\u8bbe\u5907\u6b7b\u673a\u3001\u7f51\u5173\u62e5\u5835<\/li>\n
        • \u8d85\u65f6\u515c\u5e95：\u6240\u6709\u8bbe\u5907\u6307\u4ee4\u8bbe\u7f6e 3 \u79d2\u8d85\u65f6，\u81ea\u52a8\u5224\u5b9a\u5931\u8d25\u5e76\u7559\u5b58\u65e5\u5fd7，\u9632\u6b62\u7a0b\u5e8f\u963b\u585e<\/li>\n
        • \u72b6\u6001\u5b9e\u65f6\u56de\u5199：\u6307\u4ee4\u6267\u884c\u540e\u540c\u6b65\u66f4\u65b0\u8bb0\u5fc6\u5c42\u8bbe\u5907\u72b6\u6001，\u4fdd\u8bc1\u5168\u94fe\u8def\u6570\u636e\u4e00\u81f4\u6027<\/li>\n
        • \u5e42\u7b49\u6027\u9632\u62a4：\u901a\u8fc7\u65f6\u95f4\u6233\u5b9e\u73b0\u6307\u4ee4\u53bb\u91cd，\u675c\u7edd\u91cd\u590d\u6267\u884c\u5f15\u53d1\u7684\u8bbe\u5907\u5f02\u5e38<\/li>\n<\/ul>\n

          7. \u5b8c\u6574 Agent \u4e32\u8054\u8c03\u5ea6（\u7aef\u5230\u7aef\u95ed\u73af）<\/h3>\n

          \u6574\u5408\u8bb0\u5fc6\u7ba1\u7406\u3001\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e09\u5927\u6838\u5fc3\u6a21\u5757，\u642d\u5efa\u7aef\u5230\u7aef\u8c03\u5ea6\u903b\u8f91，\u5b9e\u73b0\u4ece\u7528\u6237\u81ea\u7136\u8bed\u8a00\u8f93\u5165\u5230\u7269\u7406\u8bbe\u5907\u6267\u884c\u7684\u5168\u94fe\u8def\u95ed\u73af，\u5b8c\u6574\u8c03\u5ea6\u4ee3\u7801\u5982\u4e0b：<\/p>\n

          from<\/span> memory_manager import<\/span> IoTMemoryManager
          \nfrom<\/span> task_planner import<\/span> IoTTaskPlanner
          \nfrom<\/span> tool_caller import<\/span> IoTToolCaller<\/p>\n

          class<\/span> IoTLLMAgent<\/span>:<\/span>
          \n """IoT LLM Agent \u603b\u8c03\u5ea6\u5668"""<\/span>
          \n def<\/span> __init__<\/span>(<\/span>self)<\/span>:<\/span>
          \n self.<\/span>memory =<\/span> IoTMemoryManager(<\/span>)<\/span>
          \n self.<\/span>planner =<\/span> IoTTaskPlanner(<\/span>)<\/span>
          \n self.<\/span>caller =<\/span> IoTToolCaller(<\/span>)<\/span><\/p>\n

          def<\/span> run<\/span>(<\/span>self,<\/span> user_input:<\/span> str<\/span>)<\/span> –<\/span>><\/span> str<\/span>:<\/span>
          \n # 1. \u4efb\u52a1\u89c4\u5212<\/span>
          \n task_list =<\/span> self.<\/span>planner.<\/span>plan_task(<\/span>user_input)<\/span>
          \n if<\/span> not<\/span> task_list:<\/span>
          \n return<\/span> "\u65e0\u6709\u6548\u53ef\u6267\u884c\u8bbe\u5907\u4efb\u52a1"<\/span><\/p>\n

          # 2. \u6279\u91cf\u6267\u884c\u5de5\u5177\u8c03\u7528<\/span>
          \n exec_results =<\/span> [<\/span>]<\/span>
          \n for<\/span> task in<\/span> task_list:<\/span>
          \n res =<\/span> self.<\/span>caller.<\/span>execute_tool(<\/span>task)<\/span>
          \n exec_results.<\/span>append(<\/span>res)<\/span><\/p>\n

          # 3. \u66f4\u65b0\u8bbe\u5907\u72b6\u6001 + \u4fdd\u5b58\u957f\u671f\u8bb0\u5fc6<\/span>
          \n device_state =<\/span> str<\/span>(<\/span>{<\/span>dev:<\/span> self.<\/span>memory.<\/span>get_device_state(<\/span>dev)<\/span> for<\/span> dev in<\/span> [<\/span>"light_01"<\/span>,<\/span> "air_01"<\/span>]<\/span>}<\/span>)<\/span>
          \n self.<\/span>memory.<\/span>save_task_memory(<\/span>
          \n user_input=<\/span>user_input,<\/span>
          \n task_plan=<\/span>str<\/span>(<\/span>task_list)<\/span>,<\/span>
          \n tool_result=<\/span>str<\/span>(<\/span>exec_results)<\/span>,<\/span>
          \n device_state=<\/span>device_state
          \n )<\/span>
          \n return<\/span> f"\u4efb\u52a1\u6267\u884c\u5b8c\u6210：<\/span>{<\/span>exec_results}<\/span><\/span>"<\/span><\/span><\/p>\n

          # \u542f\u52a8 Agent<\/span>
          \nif<\/span> __name__ ==<\/span> "__main__"<\/span>:<\/span>
          \n agent =<\/span> IoTLLMAgent(<\/span>)<\/span>
          \n # \u6d4b\u8bd5\u6307\u4ee4<\/span>
          \n print<\/span>(<\/span>agent.<\/span>run(<\/span>"\u623f\u95f4\u6709\u70b9\u70ed，\u5e2e\u6211\u6253\u5f00\u7a7a\u8c03\u8c03\u523024\u5ea6，\u6253\u5f00\u706f\u5149"<\/span>)<\/span>)<\/span><\/p>\n

          8. \u843d\u5730\u8e29\u5751\u603b\u7ed3\u4e0e\u751f\u4ea7\u7ea7\u4f18\u5316\u65b9\u6848<\/h3>\n

          \u7ed3\u5408\u591a\u4e2a IoT LLM Agent \u5546\u4e1a\u5316\u843d\u5730\u7ecf\u9a8c，\u603b\u7ed3\u51fa\u5f00\u53d1\u8005\u6700\u6613\u8e29\u5751\u7684 5 \u7c7b\u6838\u5fc3\u95ee\u9898，\u540c\u6b65\u914d\u5957\u751f\u4ea7\u7ea7\u4f18\u5316\u65b9\u6848，\u89e3\u51b3\u843d\u5730\u7a33\u5b9a\u6027\u96be\u9898：<\/p>\n

          8.1 \u5751\u70b9 1：\u5927\u6a21\u578b\u57fa\u4e8e\u865a\u5047\u72b6\u6001\u63a8\u7406<\/h4>\n

          \u95ee\u9898：\u8bbe\u5907\u7269\u7406\u72b6\u6001\u5df2\u66f4\u65b0，\u4f46 Agent \u7f13\u5b58\u8bb0\u5fc6\u672a\u540c\u6b65，\u5bfc\u81f4\u5927\u6a21\u578b\u57fa\u4e8e\u65e7\u72b6\u6001\u91cd\u590d\u4e0b\u53d1\u65e0\u6548\u6307\u4ee4\u3002<\/p>\n

          \u4f18\u5316\u65b9\u6848：\u8bbe\u5907\u4e0a\u62a5\u72b6\u6001\u5b9e\u65f6\u540c\u6b65\u81f3 Redis \u7f13\u5b58，\u6bcf\u6b21\u4efb\u52a1\u89c4\u5212\u5f3a\u5236\u8bfb\u53d6\u6700\u65b0\u7f13\u5b58\u72b6\u6001，\u7981\u6b62\u9ed8\u8ba4\u63a8\u7406\u3001\u65e7\u72b6\u6001\u63a8\u7406\u3002<\/p>\n

          8.2 \u5751\u70b9 2：\u591a\u8bbe\u5907\u6307\u4ee4\u5e76\u53d1\u51b2\u7a81<\/h4>\n

          \u95ee\u9898：\u591a\u8bbe\u5907\u6307\u4ee4\u5e76\u53d1\u4e0b\u53d1，\u5f15\u53d1\u7f51\u5173\u62e5\u5835\u3001\u8bbe\u5907\u54cd\u5e94\u9519\u4e71\u3001\u6307\u4ee4\u4e22\u5931\u7b49\u95ee\u9898\u3002<\/p>\n

          \u4f18\u5316\u65b9\u6848：\u5f15\u5165\u4efb\u52a1\u961f\u5217\u673a\u5236，\u65e0\u4f9d\u8d56\u4efb\u52a1\u5e76\u884c\u6267\u884c\u3001\u5f3a\u4f9d\u8d56\u4efb\u52a1\u4e32\u884c\u6267\u884c，\u642d\u914d\u9650\u6d41\u7b56\u7565，\u7a33\u5b9a\u8c03\u5ea6\u94fe\u8def\u3002<\/p>\n

          8.3 \u5751\u70b9 3：\u79bb\u7ebf\u8bbe\u5907\u65e0\u6548\u8c03\u7528<\/h4>\n

          \u95ee\u9898：\u8bbe\u5907\u79bb\u7ebf\u540e，Agent \u6301\u7eed\u91cd\u8bd5\u8c03\u7528，\u6d6a\u8d39\u7b97\u529b\u4e0e\u7f51\u7edc\u8d44\u6e90，\u4ea7\u751f\u5927\u91cf\u65e0\u6548\u65e5\u5fd7\u3002<\/p>\n

          \u4f18\u5316\u65b9\u6848：\u65b0\u589e\u8bbe\u5907\u5728\u7ebf\u72b6\u6001\u68c0\u6d4b\u673a\u5236，\u79bb\u7ebf\u8bbe\u5907\u76f4\u63a5\u8fc7\u6ee4\u5bf9\u5e94\u4efb\u52a1，\u540c\u6b65\u8bb0\u5f55\u5f02\u5e38\u65e5\u5fd7，\u89c4\u907f\u65e0\u6548\u8c03\u5ea6\u3002<\/p>\n

          8.4 \u5751\u70b9 4：\u5927\u6a21\u578b\u8f93\u51fa\u683c\u5f0f\u9519\u4e71<\/h4>\n

          \u95ee\u9898：\u5927\u6a21\u578b\u63a8\u7406\u8f93\u51fa\u683c\u5f0f\u4e0d\u89c4\u8303，\u975e\u6807\u51c6 JSON \u5bfc\u81f4\u4efb\u52a1\u89e3\u6790\u5931\u8d25\u3001\u6267\u884c\u4e2d\u65ad\u3002<\/p>\n

          \u4f18\u5316\u65b9\u6848：\u901a\u8fc7 Prompt \u5f3a\u7ea6\u675f\u8f93\u51fa\u683c\u5f0f，\u540c\u65f6\u589e\u52a0\u683c\u5f0f\u6821\u9a8c\u4e0e\u5bb9\u9519\u89e3\u6790\u903b\u8f91，\u63d0\u5347\u94fe\u8def\u7a33\u5b9a\u6027\u3002<\/p>\n

          8.5 \u5751\u70b9 5：\u65e0\u8bb0\u5fc6\u4e2a\u6027\u5316\u80fd\u529b<\/h4>\n

          \u95ee\u9898：\u6bcf\u6b21\u4ea4\u4e92\u72ec\u7acb\u63a8\u7406，\u65e0\u6cd5\u6c89\u6dc0\u7528\u6237\u884c\u4e3a\u4e60\u60ef，\u7f3a\u4e4f\u4e2a\u6027\u5316\u667a\u80fd\u80fd\u529b\u3002<\/p>\n

          \u4f18\u5316\u65b9\u6848：\u57fa\u4e8e\u957f\u671f\u4efb\u52a1\u8bb0\u5fc6\u7edf\u8ba1\u7528\u6237\u9ad8\u9891\u64cd\u4f5c，\u81ea\u52a8\u6c89\u6dc0\u573a\u666f\u89c4\u5219，\u5b9e\u73b0\u4e2a\u6027\u5316\u667a\u80fd\u8c03\u5ea6\u3002<\/p>\n

          9. \u603b\u7ed3\u4e0e\u540e\u7eed\u62d3\u5c55\u65b9\u5411<\/h3>\n

          \u672c\u6587\u4ece\u96f6\u642d\u5efa\u4e86\u4e00\u5957\u5b8c\u6574\u7684\u751f\u4ea7\u7ea7 IoT LLM Agent，\u9488\u5bf9\u6027\u89e3\u51b3\u901a\u7528 Agent \u4e0d\u9002\u914d\u7269\u7406\u7269\u8054\u7f51\u573a\u666f\u7684\u6838\u5fc3\u75db\u70b9\u3002\u901a\u8fc7\u573a\u666f\u5316\u8bb0\u5fc6\u7ba1\u7406\u3001\u786c\u4ef6\u7ea6\u675f\u4efb\u52a1\u89c4\u5212\u3001\u6807\u51c6\u5316\u8bbe\u5907\u5de5\u5177\u8c03\u7528\u4e09\u5927\u6838\u5fc3\u6a21\u5757，\u5b9e\u73b0\u81ea\u7136\u8bed\u8a00\u5230\u7269\u7406\u8bbe\u5907\u64cd\u4f5c\u7684\u5b89\u5168\u3001\u7cbe\u51c6\u3001\u9ad8\u6548\u95ed\u73af\u3002<\/p>\n

          \u8be5\u67b6\u6784\u843d\u5730\u6210\u672c\u4f4e\u3001\u7a33\u5b9a\u6027\u5f3a\u3001\u9002\u914d\u6027\u5e7f，\u53ef\u76f4\u63a5\u590d\u7528\u81f3\u667a\u80fd\u5bb6\u5c45\u3001\u5de5\u4e1a\u7269\u8054\u7f51\u3001\u667a\u6167\u529e\u516c\u3001\u667a\u6167\u56ed\u533a\u7b49\u573a\u666f，\u652f\u6301\u8fb9\u7f18\u672c\u5730\u90e8\u7f72，\u65e0\u9700\u4f9d\u8d56\u4e91\u7aef\u5927\u6a21\u578b，\u5b8c\u7f8e\u5951\u5408\u7269\u8054\u7f51\u4f4e\u5ef6\u8fdf\u3001\u9ad8\u53ef\u9760\u7684\u6838\u5fc3\u8bc9\u6c42\u3002<\/p>\n

          \u540e\u7eed\u8fdb\u9636\u62d3\u5c55\u65b9\u5411：<\/p>\n

            \n
          • \u5f15\u5165\u5411\u91cf\u8bb0\u5fc6，\u5b9e\u73b0\u76f8\u4f3c\u573a\u666f\u667a\u80fd\u590d\u7528，\u5f3a\u5316 Agent \u4e2a\u6027\u5316\u4ea4\u4e92\u80fd\u529b<\/li>\n
          • \u642d\u5efa\u591a Agent \u534f\u540c\u67b6\u6784，\u62c6\u5206\u611f\u77e5\u3001\u89c4\u5212\u3001\u6267\u884c\u5b50 Agent，\u9002\u914d\u590d\u6742\u5de5\u4e1a IoT \u573a\u666f<\/li>\n
          • \u63a5\u5165\u4f20\u611f\u5668\u6570\u636e\u5f02\u5e38\u9884\u8b66，\u5b9e\u73b0\u8bbe\u5907\u4e3b\u52a8\u667a\u80fd\u8c03\u5ea6，\u6253\u7834\u88ab\u52a8\u54cd\u5e94\u6a21\u5f0f<\/li>\n
          • \u57fa\u4e8e IoT \u4e13\u5c5e\u6570\u636e\u96c6\u8f7b\u91cf\u5316\u5fae\u8c03\u6a21\u578b，\u8fdb\u4e00\u6b65\u63d0\u5347\u4efb\u52a1\u89c4\u5212\u7cbe\u5ea6，\u6839\u9664\u65e0\u6548\u6307\u4ee4<\/li>\n<\/ul>\n

            \u5199\u5728\u6700\u540e：\u5927\u6a21\u578b + \u7269\u8054\u7f51\u5e76\u975e\u7b80\u5355\u6280\u672f\u62fc\u63a5，\u800c\u662f AI \u8bed\u4e49\u667a\u80fd\u4e0e\u7269\u7406\u4e16\u754c\u89c4\u5219\u7684\u6df1\u5ea6\u878d\u5408\u3002\u8131\u79bb\u786c\u4ef6\u7ea6\u675f\u4e0e\u573a\u666f\u89c4\u8303\u7684 Agent \u53ea\u662f\u7a7a\u4e2d\u697c\u9601，\u552f\u6709\u8d34\u5408 IoT \u573a\u666f\u7279\u6027，\u5b8c\u5584\u8bb0\u5fc6\u3001\u89c4\u5212\u3001\u8c03\u5ea6\u5168\u94fe\u8def\u80fd\u529b，\u624d\u80fd\u5b9e\u73b0\u771f\u6b63\u53ef\u5546\u7528\u3001\u9ad8\u7a33\u5b9a\u7684\u667a\u80fd\u7269\u8054\u7f51\u843d\u5730\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"

            1. \u5f15\u8a00
            \n\u5f53\u524d LLM Agent \u5927\u591a\u805a\u7126\u7eaf\u8f6f\u4ef6\u573a\u666f，\u9002\u914d\u529e\u516c\u81ea\u52a8\u5316\u3001\u77e5\u8bc6\u5e93\u95ee\u7b54\u3001\u6570\u5b57\u4eba\u4ea4\u4e92\u7b49\u8f7b\u91cf\u5316\u9700\u6c42\u3002\u4e0d\u5c11\u5f00\u53d1\u8005\u76f4\u63a5\u5957\u7528\u901a\u7528 Agent \u67b6\u6784\u843d\u5730 IoT \u573a\u666f，\u6700\u7ec8\u9891\u53d1\u6307\u4ee4\u65e0\u6548\u3001\u8bbe\u5907\u51b2\u7a81\u3001\u72b6\u6001\u9519\u4e71\u3001\u91cd\u590d\u6267\u884c\u3001\u79bb\u7ebf\u5931\u6548\u7b49\u95ee\u9898，\u5bfc\u81f4\u9879\u76ee\u843d\u5730\u5361\u58f3\u3002
            \nIoT Agent \u7684\u6838\u5fc3\u7279\u8d28\u662f\u7269\u7406\u4e16\u754c\u95ed\u73af\u4ea4\u4e92，\u4e0e\u7eaf\u8f6f\u4ef6 Agent \u5b58\u5728\u672c\u8d28\u5dee\u5f02：\u7269\u7406\u8bbe\u5907\u5b58\u5728\u786c\u4ef6\u9650\u5236\u3001\u72b6\u6001\u5ef6\u8fdf\u3001\u64cd\u4f5c\u4e0d\u53ef\u9006\u3001\u7f51\u7edc\u6ce2\u52a8\u3001\u6307\u4ee4\u5e26\u5b9e\u4f53\u526f\u4f5c\u7528<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[230,190,220],"topic":[],"class_list":["post-110470","post","type-post","status-publish","format-standard","hentry","category-server","tag-php","tag-190","tag-220"],"yoast_head":"\n\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/www.wsisp.com\/helps\/110470.html\" \/>\n<meta property=\"og:locale\" content=\"zh_CN\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"og:description\" content=\"1. \u5f15\u8a00 \u5f53\u524d LLM Agent \u5927\u591a\u805a\u7126\u7eaf\u8f6f\u4ef6\u573a\u666f，\u9002\u914d\u529e\u516c\u81ea\u52a8\u5316\u3001\u77e5\u8bc6\u5e93\u95ee\u7b54\u3001\u6570\u5b57\u4eba\u4ea4\u4e92\u7b49\u8f7b\u91cf\u5316\u9700\u6c42\u3002\u4e0d\u5c11\u5f00\u53d1\u8005\u76f4\u63a5\u5957\u7528\u901a\u7528 Agent \u67b6\u6784\u843d\u5730 IoT \u573a\u666f，\u6700\u7ec8\u9891\u53d1\u6307\u4ee4\u65e0\u6548\u3001\u8bbe\u5907\u51b2\u7a81\u3001\u72b6\u6001\u9519\u4e71\u3001\u91cd\u590d\u6267\u884c\u3001\u79bb\u7ebf\u5931\u6548\u7b49\u95ee\u9898，\u5bfc\u81f4\u9879\u76ee\u843d\u5730\u5361\u58f3\u3002 IoT Agent \u7684\u6838\u5fc3\u7279\u8d28\u662f\u7269\u7406\u4e16\u754c\u95ed\u73af\u4ea4\u4e92，\u4e0e\u7eaf\u8f6f\u4ef6 Agent \u5b58\u5728\u672c\u8d28\u5dee\u5f02：\u7269\u7406\u8bbe\u5907\u5b58\u5728\u786c\u4ef6\u9650\u5236\u3001\u72b6\u6001\u5ef6\u8fdf\u3001\u64cd\u4f5c\u4e0d\u53ef\u9006\u3001\u7f51\u7edc\u6ce2\u52a8\u3001\u6307\u4ee4\u5e26\u5b9e\u4f53\u526f\u4f5c\u7528\" \/>\n<meta property=\"og:url\" content=\"https:\/\/www.wsisp.com\/helps\/110470.html\" \/>\n<meta property=\"og:site_name\" content=\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\" \/>\n<meta property=\"article:published_time\" content=\"2026-09-28T10:24:50+00:00\" \/>\n<meta name=\"author\" content=\"admin\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<meta name=\"twitter:label1\" content=\"\u4f5c\u8005\" \/>\n\t<meta name=\"twitter:data1\" content=\"admin\" \/>\n\t<meta name=\"twitter:label2\" content=\"\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4\" \/>\n\t<meta name=\"twitter:data2\" content=\"7 \u5206\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\/\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/110470.html\",\"url\":\"https:\/\/www.wsisp.com\/helps\/110470.html\",\"name\":\"\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"isPartOf\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\"},\"datePublished\":\"2026-09-28T10:24:50+00:00\",\"dateModified\":\"2026-09-28T10:24:50+00:00\",\"author\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\"},\"breadcrumb\":{\"@id\":\"https:\/\/www.wsisp.com\/helps\/110470.html#breadcrumb\"},\"inLanguage\":\"zh-Hans\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\/\/www.wsisp.com\/helps\/110470.html\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/110470.html#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"\u9996\u9875\",\"item\":\"https:\/\/www.wsisp.com\/helps\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#website\",\"url\":\"https:\/\/www.wsisp.com\/helps\/\",\"name\":\"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3\",\"description\":\"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b\",\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}\"},\"query-input\":\"required name=search_term_string\"}],\"inLanguage\":\"zh-Hans\"},{\"@type\":\"Person\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41\",\"name\":\"admin\",\"image\":{\"@type\":\"ImageObject\",\"inLanguage\":\"zh-Hans\",\"@id\":\"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/\",\"url\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"contentUrl\":\"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery\",\"caption\":\"admin\"},\"sameAs\":[\"http:\/\/wp.wsisp.com\"],\"url\":\"https:\/\/www.wsisp.com\/helps\/author\/admin\"}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/www.wsisp.com\/helps\/110470.html","og_locale":"zh_CN","og_type":"article","og_title":"\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","og_description":"1. \u5f15\u8a00 \u5f53\u524d LLM Agent \u5927\u591a\u805a\u7126\u7eaf\u8f6f\u4ef6\u573a\u666f，\u9002\u914d\u529e\u516c\u81ea\u52a8\u5316\u3001\u77e5\u8bc6\u5e93\u95ee\u7b54\u3001\u6570\u5b57\u4eba\u4ea4\u4e92\u7b49\u8f7b\u91cf\u5316\u9700\u6c42\u3002\u4e0d\u5c11\u5f00\u53d1\u8005\u76f4\u63a5\u5957\u7528\u901a\u7528 Agent \u67b6\u6784\u843d\u5730 IoT \u573a\u666f，\u6700\u7ec8\u9891\u53d1\u6307\u4ee4\u65e0\u6548\u3001\u8bbe\u5907\u51b2\u7a81\u3001\u72b6\u6001\u9519\u4e71\u3001\u91cd\u590d\u6267\u884c\u3001\u79bb\u7ebf\u5931\u6548\u7b49\u95ee\u9898，\u5bfc\u81f4\u9879\u76ee\u843d\u5730\u5361\u58f3\u3002 IoT Agent \u7684\u6838\u5fc3\u7279\u8d28\u662f\u7269\u7406\u4e16\u754c\u95ed\u73af\u4ea4\u4e92，\u4e0e\u7eaf\u8f6f\u4ef6 Agent \u5b58\u5728\u672c\u8d28\u5dee\u5f02：\u7269\u7406\u8bbe\u5907\u5b58\u5728\u786c\u4ef6\u9650\u5236\u3001\u72b6\u6001\u5ef6\u8fdf\u3001\u64cd\u4f5c\u4e0d\u53ef\u9006\u3001\u7f51\u7edc\u6ce2\u52a8\u3001\u6307\u4ee4\u5e26\u5b9e\u4f53\u526f\u4f5c\u7528","og_url":"https:\/\/www.wsisp.com\/helps\/110470.html","og_site_name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","article_published_time":"2026-09-28T10:24:50+00:00","author":"admin","twitter_card":"summary_large_image","twitter_misc":{"\u4f5c\u8005":"admin","\u9884\u8ba1\u9605\u8bfb\u65f6\u95f4":"7 \u5206"},"schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/www.wsisp.com\/helps\/110470.html","url":"https:\/\/www.wsisp.com\/helps\/110470.html","name":"\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218 - \u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","isPartOf":{"@id":"https:\/\/www.wsisp.com\/helps\/#website"},"datePublished":"2026-09-28T10:24:50+00:00","dateModified":"2026-09-28T10:24:50+00:00","author":{"@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41"},"breadcrumb":{"@id":"https:\/\/www.wsisp.com\/helps\/110470.html#breadcrumb"},"inLanguage":"zh-Hans","potentialAction":[{"@type":"ReadAction","target":["https:\/\/www.wsisp.com\/helps\/110470.html"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/www.wsisp.com\/helps\/110470.html#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"\u9996\u9875","item":"https:\/\/www.wsisp.com\/helps"},{"@type":"ListItem","position":2,"name":"\u4ece 0 \u5230 1 \u6784\u5efa IoT \u573a\u666f\u4e0b\u7684 LLM Agent\uff1a\u4efb\u52a1\u89c4\u5212\u3001\u5de5\u5177\u8c03\u7528\u4e0e\u8bb0\u5fc6\u7ba1\u7406\u5b9e\u6218"}]},{"@type":"WebSite","@id":"https:\/\/www.wsisp.com\/helps\/#website","url":"https:\/\/www.wsisp.com\/helps\/","name":"\u7f51\u7855\u4e92\u8054\u5e2e\u52a9\u4e2d\u5fc3","description":"\u9999\u6e2f\u670d\u52a1\u5668_\u9999\u6e2f\u4e91\u670d\u52a1\u5668\u8d44\u8baf_\u670d\u52a1\u5668\u5e2e\u52a9\u6587\u6863_\u670d\u52a1\u5668\u6559\u7a0b","potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/www.wsisp.com\/helps\/?s={search_term_string}"},"query-input":"required name=search_term_string"}],"inLanguage":"zh-Hans"},{"@type":"Person","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/358e386c577a3ab51c4493330a20ad41","name":"admin","image":{"@type":"ImageObject","inLanguage":"zh-Hans","@id":"https:\/\/www.wsisp.com\/helps\/#\/schema\/person\/image\/","url":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","contentUrl":"https:\/\/gravatar.wp-china-yes.net\/avatar\/?s=96&d=mystery","caption":"admin"},"sameAs":["http:\/\/wp.wsisp.com"],"url":"https:\/\/www.wsisp.com\/helps\/author\/admin"}]}},"_links":{"self":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/110470","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/comments?post=110470"}],"version-history":[{"count":0,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/posts\/110470\/revisions"}],"wp:attachment":[{"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/media?parent=110470"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/categories?post=110470"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/tags?post=110470"},{"taxonomy":"topic","embeddable":true,"href":"https:\/\/www.wsisp.com\/helps\/wp-json\/wp\/v2\/topic?post=110470"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}