站在AI和IoT的十字路口，如何喚醒萬物？

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"日前，百度對外分享了其在物聯網的智能化道路上的最新思考、實踐和總結。6月下旬，2021百度智能雲“雲智技術論壇”智能物聯網專場在蘇州舉辦。會上，百度智能雲技術專家及合作伙伴共同探討了物聯網在智能化方向的前沿基礎及應用技術，分享了打造AIoT最佳應用實踐的經驗。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"智聯網思維"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在百度看來，當前“萬物互聯”的發展關鍵詞是“連接”、“對話”和“智能”。“連接”層面需要具備億級規模設備連接的能力支持，而低延時連接技術在無人車、遠程駕駛等場景下，也面臨着新的技術挑戰。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"“對話”意在賦予萬物對話能力，涉及到數字化控制、語音、視覺和虛擬現實等技術。“智能”則是在連接和對話能力的基礎上，將AI、大數據處理和分析等技術能力，深入融合到各行各業的應用場景。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"“我們的思維方式要從傳統的物聯網思維轉到智聯網思維，考慮的不再是一個個的傳感器以及把數據放在一起，而是如何把智能做深入融合，給行業帶來新的機會和方法。”百度副總裁馬傑表示，希望用AI來給IoT拓展新的場景，增加網聯化和數據價值，完成用AI喚醒並賦能萬物的使命。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"考慮到穩定性和功耗以及成本等因素，現在每一個設備都需要做得簡單可靠、少耗電、少算力，那如何在低算力、低功耗的環境下，提供AI能力？對此，馬傑舉了兩個例子："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"1.利用邊緣能力，用低功耗的方式跟最近的邊緣盒子連接，由邊緣盒子提供技術能力。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"2.在雲端提供完整的基於百度雲架構的遠程算力。遠程算力提供完整的用戶交互、計算能力，在需要使用的時候把它跟一個前端設備連接，讓它可以在不需要使用的時候完全釋放掉。在雲端跑的是安卓系統，相當於在前端每一個小傳感器上同時具有完整的安卓系統和強大算力的CPU，但不用的時候又可以釋放掉。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"馬傑指出，除此之外百度還有不同的智慧萬物解決方案，而這些都是基於百度智能雲所做的積累，包括AI操作系統PP飛槳、AI芯片百度崑崙以及鴻鵠語音交互芯片等等。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"物聯網平臺價值的重新發現"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"一個被不斷思考的問題是：在當前的碎片化場景下，物聯網平臺的價值度如何？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"百度智能雲物聯網部總經理曹楊認爲，從現在的結果來看，基於連接爲主的物聯網平臺，除了終端連接數不斷增長，但經濟價值、技術競爭價值及支撐應用場景上的價值，相對來講還是沒有發揮非常好的狀態，包括最終支撐到應用落地、數據價值挖掘，從現在傳統的物聯網平臺角度來講，其實並沒有涉足太多，更多是聚焦在連接本身，把連接的穩定性、質量、延時做到足夠高，但這遠遠不夠。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"智能化在物聯網平臺中已是必然趨勢。物聯網平臺從設備的管理、設備接入、數據流轉到數據價值的挖掘，這是物聯網平臺從開始到結束的完整過程。在數據流傳的過程中，大的流程沒有出現大的變化，但數據的形態及中間過程中不僅僅只是停留在數據分析，而是在多模數據交互上催生出新的應用場景，這極大地擴大了物聯網領域應用和技術上的挑戰。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在百度看來，這個技術挑戰可以總結爲從最初的做“設備上雲連接”變成做“邊緣雲協同”，對連接的時效性、數據的私密性都提出了新的要求。如果是語音智能家電場景，對端的要求則更高。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"曹楊稱，現階段，端邊雲協同框架已經比較成熟，並且在一些典型應用場景中做了最佳的實踐落地。當天，他介紹了天工AIoT平臺2.0，"},{"type":"link","attrs":{"href":"https:\/\/cloud.baidu.com\/solution\/iot\/index.html","title":"xxx","type":null},"content":[{"type":"text","text":"天工AIoT平臺"}]},{"type":"text","text":"以“端邊雲”協同架構支撐多種類型數據價值的分析和交互。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.infoq.cn\/resource\/image\/ba\/8f\/baa37570e98510cc5595d9da8f65048f.png","alt":null,"title":"","style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":"","fromPaste":false,"pastePass":false}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":"center","origin":null},"content":[{"type":"text","marks":[{"type":"size","attrs":{"size":10}}],"text":"天工AIoT平臺2.0全景圖"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"全景圖囊括了從最底層的智能硬件、芯片、終端模組、智能終端，到邊緣端、平臺端、能力端（也可以稱之爲算力端），整個平臺層、能力層較最初有了極大的擴充。平臺有了，應用牽引是所有技術的一個最終落腳點。而AIoT能力正好能夠契合工業、能源、園區安全、家電等等產業升級的訴求，彼此可以相互促進往前發展。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"立足於天工AIoT平臺，百度亦希望用開放的態度把技術輸出，將“端—邊核心技術框架”開源，一個是"},{"type":"link","attrs":{"href":"https:\/\/github.com\/baetyl","title":"xxx","type":null},"content":[{"type":"text","text":"Baetyl邊緣計算開源框架"}]},{"type":"text","text":"，另一個是結合鴻鵠語音芯片的開源框架——"},{"type":"link","attrs":{"href":"https:\/\/github.com\/baidu\/du1906_esp","title":"xxx","type":null},"content":[{"type":"text","text":"百度Du1906-Korvo語音設備端開源框架"}]},{"type":"text","text":"。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"天工平臺技術解讀"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"百度智能雲物聯網主架構師李樂丁表示，天工AIoT平臺是一個位於數據算法和基礎設施交匯點的PaaS平臺。“我們一方面向下充分挖掘百度智能雲在IaaS、邊緣和異構計算方面的強大算力能力，另外一方面向上和合作夥伴一起拓展各種垂直領域解決方案。在技術上一方面爲各種物聯網設備終端提供STK，簡化開發模型，另外一方面也提供設備互聯互通中心平臺，以及爲開發者提供多種多樣能夠幫助他快速開發應用程序的中間件。”"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"據瞭解，天工AIoT平臺是國內最早實現分佈式雲原生和多租戶的IoT平臺。每天的設備連接數超過兩千萬，收發超過7億條消息，每一條都要在不到10毫秒的時間內得到處理。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"除了擁有海量數據接入能力，天工還支持更多協議。天工平臺已經開始支持大量MQTT5.0先進特性，並尤其着重去加強了應用程序在負載均衡和流量控制方面的能力。除了MQTT，還進一步增加了對COAP和HTTP兩種協議的支持。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"協議之上還增加了對於更多數據種類的支持，不僅僅可以在MQTT裏傳輸文本、數字，還可以在MQTT協議上直接傳輸語音數據，在一個智能終端上簡化網絡編程接口，提升帶寬利用率及延遲降低。舉個例子，將天工AIoT平臺和百度的Apollo自動駕駛結合，實現了道路的智能化。在路測收集過來的圖片、視頻信息及傳感器信息，將由天工平臺傳輸和實時進行AI處理，處理之後得到的路面信息將被快速傳送到周邊的自動駕駛汽車上，那麼車可以更早、更遠地看到路面狀況，從而改善駕駛安全，提升道路通行效率。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"通過運用智能化手段，天工AIoT平臺對時間數據和空間數據進行結合處理分析，解決數據孤島難題，進行有效的數據治理及數據共享，提高運營效率；爲了提供更優的編程模式，天工AIoT平臺以數字單元串聯能力構造智能應用，實現可視化無代碼編程。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"今天做一個智能物聯網應用，面臨的主要困難是設備的多樣性、場景的分散性及技術體系的碎片化，爲此百度提出了“數字單元”。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.infoq.cn\/resource\/image\/0d\/5e\/0d7f221caa609bfd100f191357b1905e.jpeg","alt":null,"title":"","style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":"","fromPaste":false,"pastePass":false}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"據李樂丁介紹，數字單元是基於數字孿生能力進一步抽象而來，每一個終端設備、每一個算法、每一個計算節點都被抽象成爲一種標準的數字單元，而所有東西都將共享同樣的行爲，它們都有標準化的輸入、標準化的輸出及標準化的屬性接口。因爲標準化，所以開發者只需要很簡單地把所需要的數字單元放在一起，然後畫出連接關係，就可以得到智能應用。背後不同單元面臨的差異性、工程部署問題、監控問題、運維問題等等一切的複雜性都交給平臺。最後百度會通過技術和跟合作伙伴的共同努力解決這些問題，讓開發者得到釋放。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.infoq.cn\/resource\/image\/de\/a7\/ded0dbf822ca311dd234f4e387151ba7.jpeg","alt":null,"title":"","style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":"","fromPaste":false,"pastePass":false}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"目前該工具在雲端可以直接使用。以上圖爲例，左側列出了開發者目前手中可以得到的東西，比如他可以接入流媒體視頻或者進行各種各樣的AI處理，而且很多AI是經由合作伙伴已經做好的相應場景化。開發者只需要把這些東西拖動到面板中，然後讓它們之間建立數據連接，畫出一條線，畫完之後平臺會自動將這張圖編譯成物聯網應用，自動將這張圖下發到合適節點，無論是雲端，還是邊緣側、設備終端，背後平臺會自動保證它們的數據連通性運行。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"“我們會支持更多的數據協議，在數據協議上支持語音、圖像等等更多的數據內容，我們會充分利用分佈式、雲原生及協議高級特性支持數據的海量爆發，有了大量的數據我們會提供時空大數據、數據孿生等等先進的算法能力，用時空大數據解決空間和時間的結合問題，用數字孿生讓所有的設備變的可知、可控，最後我們會用全新的應用開發平臺讓物聯網應用從複雜變到簡單。”李樂丁總結道。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"或許正如馬傑所述，我們已經站在IoT和AI融合的關鍵路口，IoT的網聯化和智能化，將爲IoT發展帶來新價值。讓我們拭目以待。"}]}]}