︿
Top

自動駕駛車與深度學習專利佈局:新創公司Drive.ai、Zoox

瀏覽次數:8087| 歡迎推文: facebook twitter wechat Linked

科技產業資訊室 (iKnow) - Stanley 發表於 2017年6月1日
facebook twitter wechat twitter
表1、自動駕駛控制能力SAE分級

2015年,在美國西岸加州的矽谷,有兩家系出同門的新創公司巧合地在同一年被成立,一家取名為Drive.ai,另一家則為Zoox Inc.,兩家新創公司的創辦人都曾經在Stanford University長期從事與車輛的自動駕駛有關的人工智慧技術的研發工作,進一步地,兩家公司也都將現階段「自動駕駛車輛」的自動化操控水準或目標設定在符合2014年版SAE J3016標準所提出的Level 4等級(表1),且兩家公司不約而同地都強調為車輛裝備『深度學習能力(capability of deep learning)』才是車輛駕駛自動化最根本的驅動力所在。

Drive.ai公司的聯合創始人Carol Reiley提出以下看法:自動駕駛車輛應具備的技術至少必須包含有「資料感測(data sensing)」、「物件標記(object annotation)」、「深度學習演算法(deep learning algorithm)」及「車輛方向盤、油門與煞車等對應的行車操控」等,此外Reiley強調Drive.ai的核心技術在於使車輛具有如同人類駕駛一般的腦袋,可透過不同駕駛場景的練習而不斷地累積駕駛經驗,再從過去的經驗中衍生出得以因應未來面對新的駕駛場景時的反應能力,相較於以規則為基礎(rule-based)的自動駕駛系統而言,Drive.ai深度學習的能力更能適應各種不同複雜場景,且毋須依賴成本昂貴的光達(LiDAR)、影像擷取設備、以及精準的3D圖資資料庫來將不斷變化的大量行車環境資料窮盡不斷地搜集、處理。
(資料來源:Reference 3, 4, 5)
 
可惜的是,Drive.ai從2015年成立以來,目前暫時沒能找到任何關於Drive.ai自己的專利申請在案,也許是提出申請但尚未公開,又或者是,關於Drive.ai深度學習演算法的內容被以營業秘密的方式加以保密。
 
為了與Drive.ai比較,現以Zoox Inc.為例進行說明:相較於Drive.ai仍持續地在尋求合作夥伴,Zoox Inc.背後有Google、Tesla及車廠的強力金援,使得Zoox Inc.自2015年成立以來,在同年11月已密集且大量地提出專利申請在案,統計約有二十餘件美國發明專利,如下表2彙整內容,依據SAE J3016標準所提出的、不同等級所需的四種自動駕駛控制能力F1至F4,經分類後可以發現,Zoox Inc.將其研發資源及專利佈局的重心放在F2(人/車/障礙物等行車環境監測)與F4(系統就各種路況自適應能力)兩大部份,由此足見Zoox Inc.在自動駕駛車輛的深度學習技術研發上是十分積極地投入的。(資料來源:Reference 6)(1050字;圖1;表2)

[註]自動駕駛分成5個等級:
Level 0 完全人工駕駛算是;
Level 1一項或多項是自動定速功能單獨運行;
Level 2自動倒車等多項功能自動化;
Level 3可完成部分駕駛任務,在一定條件下監控路面情況,司機可以不用操作,但需要隨時準備好接管駕駛;
Level 4在特定環境中,車輛可以完成所有駕駛和環境監測功能;
Level 5是最終目標:完全自動化。
要實現全自動化的無人駕駛,除了要有Level 5的智慧駕駛技術之外,更重要的是路權跟法令的配合,才能提供適合無人車上路的場域。


表2、將“Zoox Inc.”於2015至2017年期間申請/公開/公告的專利申請案,依「自動駕駛控制能力的F1至F4」進行分類
SAE分級 申請案號碼 發明名稱
F1 US 20170120753 A1 (14/757,015) Independent steering, power torque control and transfer in autonomous vehicles
US 20170124781 A1 (14/756,996) Calibration for autonomous vehicle operation
F2 US 20170120814 A1 (14/756,993) Method for robotic vehicle communication with an external environment via acoustic beam forming
US 20170123428 A1 (14/756,991) Sensor-based object-detection optimization for autonomous vehicles
US 20170120804 A1 (14/932,948) Active lighting control for communicating a state of an autonomous vehicle to entities in a surrounding environment
US 20170120902 A1 (14/932,952) Resilient safety system for a robotic vehicle
US 20170124476 A1 (14/932,940) Automated extraction of semantic information to enhance incremental mapping modifications for robotic vehicles
US 20170120803 A1 (14/756,994) System of configuring active lighting to indicate directionality of an autonomous vehicle
US 9494940 B1 (14/932,958) Quadrant configuration of robotic vehicles
US 9517767 B1 (14/932,954) Internal safety systems for robotic vehicles
F3 Zoox Inc.目前查無任何關於倒車/停車/後方障礙物偵測等技術的專利申請案
F4 US 20170123419 A1/
US 9632502 B1 (14/933,602)
Machine-learning systems and techniques to optimize teleoperation and/or planner decisions
US 20170123422 A1 (14/933,706) Interactive autonomous vehicle command controller
US 20170126810 A1 (14/933,665) Software application and logic to modify configuration of an autonomous vehicle
US 20170123429 A1 (14/756,992) Adaptive autonomous vehicle planner logic
US 20170120904 A1 (14/932,962) Robotic vehicle active safety systems and methods
US 20170123421 A1 (14/756,995) Coordination of dispatching and maintaining fleet of autonomous vehicles
US 9606539 B1 (14/932,959) Autonomous vehicle fleet service and system
US 9630619 B1 (14/932,962) Robotic vehicle active safety systems and methods
US 9612123 B1 (14/932,963) Adaptive mapping to navigate autonomous vehicles responsive to physical environment changes
US 9507346 B1 (14/932,966) Teleoperation system and method for trajectory modification of autonomous vehicles
註1. 上述表2已列之各個專利申請案,其申請日(filing date)或優先權日(priority date)皆為同一日「2015/11/4」。
註2. 檢索記錄-Google Patent KWs: inassignee:"Zoox Inc."
 
 
References:
  1. https://www.sae.org/misc/pdfs/automated_driving.pdf
  2. http://www.techrepublic.com/article/autonomous-driving-levels-0-to-5-understanding-the-differences/
  3. https://www.drive.ai/team
  4. http://spectrum.ieee.org/cars-that-think/transportation/self-driving/how-driveai-is-mastering-autonomous-driving-with-deep-learning
  5. https://techxplore.com/news/2016-04-driveai-deep-smarts-autonomous-cars.html
  6. http://fortune.com/2016/03/23/zoox-self-driving-permit/
 
--Backup Information--
Note 1:
Drive.ai is a Silicon Valley start-up founded by former lab mates out of Stanford University's Artificial Intelligence Lab. They brought to their startup knowhow in valuable areas: natural language processing, computer vision and autonomous driving.

Reiley explained how their approach matters: "We're solving the problem of a self driving car by using deep learning for the full autonomous integrated driving stack—from perception, to motion planning, to controls—as opposed to just bits and pieces like other companies have been using for autonomy. We're using an integrated architecture to create a more seamless approach."

Ackerman said Drive.ai is the 13th company to be granted a license to test autonomous vehicles on public roads in California.
(https://techxplore.com/news/2016-04-driveai-deep-smarts-autonomous-cars.html)
 
Note 2:
Zoox, a startup with aspirations for an Uber-like service, has received a permit from California to begin testing self-driving vehicles on the state's public roads.
The company will become the 12th to receive a permit from the state, and joins an illustrious list of autonomous vehicle-dreaming companies to have received permission from California, according to Bloomberg.
(http://fortune.com/2016/03/23/zoox-self-driving-permit/)


 

本站相關文章:

  1. 全球30家公司投入無人駕駛車研發
  2. 2035年全球自動駕駛車可達2100萬輛
  3. 美國發佈全球第一部自動駕駛車政策指南
  4. 無人駕駛車依舊存在著車禍的風險
  5. 自動駕駛汽車的普及可能帶來的影響

 

 
歡迎來粉絲團按讚!
--------------------------------------------------------------------------------------------------------------------------------------------
【聲明】
1.科技產業資訊室刊載此文不代表同意其說法或描述,僅為提供更多訊息,也不構成任何投資建議。
2.著作權所有,非經本網站書面授權同意不得將本文以任何形式修改、複製、儲存、傳播或轉載,本中心保留一切法律追訴權利。