透過您的圖書館登入
IP:3.145.178.106
  • 學位論文

機器人鋪面檢測之運動模式

Motion Planning Method for Robotics Pavement Inspection

指導教授 : 康仕仲

摘要


本論文發展以自主式機器人進行鋪面檢測之運動模式,其運動模式分為兩大類,其一由人工檢測方式所設計之「全面檢測模式」,另由工程中抽樣檢測概念所設計之「抽樣檢測模式」,再將運動模式由構想層面設計成控制演算法,並以CSharp程式開發出軟體模擬測試環境,用以確保演算法符合運動模式之設計,最後將演算法於硬體實現,建構自主式機器人以印證鋪面檢測運動模式。 在鋪面檢測運動模式之軟體演算法部分,共分為運動模式設計與軟體模擬部分,我們由現行檢測的模式發展為全面檢測模式,其中包含(1)縱向檢測,(2)橫向檢測,及工程抽樣檢測的概念發展出抽樣檢測模式,包含(1)隨機移動檢測,(2)方格檢測等,在透過自行開發之軟體模擬結果中發現,全面檢測模式與抽樣檢測模式有應用上之互補性。以檢測完整道路之時間為考量,全面檢測模式可應用於需要檢測道路完整的資訊,如進行新建道路驗收等,而不同於全面檢測模式以循序漸進的方式進行檢測,抽樣檢測模式採用隨機的行走方式,故若相同時間裡,受檢測路面與未受測路面的平均涵蓋度為考量,抽樣檢測模式可取得均佈的道路資訊。 建置自主式機器人的部分,為避免機器人針對單一問題來設計,導致機器人若要重新處理不同問題,必需花費許多時間與人力成本來改變硬體架構與演算核心。本文之機器人改變傳統的作法,建構出具彈性的軟硬體架構。硬體部分將採用Basic Stamp II單晶片作為訊號處理,運動方式使用全向輪,並利用機器人全向輪的特性,增加兩個特殊運動模式(1)迴避障礙物、(2)定點密集檢測以應用在檢測鋪面破壞點與躲避障礙物時之應用。此機器人在機構採用三層式架構,分別為(1)邏輯運算層、(2)訊號感知層與(3)運動控制層,分別獨立,增加硬體擴充與可修改性。軟體架構則使用Microsoft Robotics Developer Studio(MSRDS),它採用的分散式服務系統設計也基於賦予本研究之機器人更多的彈性,使得當需求改變時,可輕易的變動軟硬體來切合更多的研究主題。 在本論文之運動檢測應用層面提出取得道路真實剖面值的方式,以解決並擴展運動模式於現行檢測中之使用限制與應用範圍,最後在自主式機器人建置的部分,驗證驗算法符合設計,於鋪面檢測運動模式與建構自主式機器人的相關經驗,提供未來相關性研究之參考。

並列摘要


We propose novel pavement inspection motion modes for autonomous robots which are generalized into complete inspection modes and sampling inspection modes. Based on these two complementary modes, we are able to provide a robust and efficient pavement inspection algorithm. In details, complete inspection composes of (i) longitudinal mode and (ii) transversal mode. They are derived from tradition manual pavement inspection methods. Meanwhile, we have sampling inspection compose of (iii) random mode and (iv) grid mode which are derived from sampling inspection methods in the engineering. Then, we carry out data analysis from software simulation. The results show that complete inspection modes are faster than sampling inspection modes when the overall pavement is to be inspected. However, for complete inspection modes where there is a time constraint, sampling inspection modes will be able to cover more road information. The hardware platform is divided into three layers: a motion and sensing module, a control module, and a logic module. Developers can add, remove, or replace modules of all layers with compatible ones at will. The Basic Stamp II (BS-II) signal and control chip is used for the control module. For software, the packaged software Microsoft Robotics Developer Studio (MSRDS) is used to allow for re-use. MSRDS provides Concurrency and Coordination Runtime (CCR) and Decentralized System Services (DSS) for autonomous tasks. The overall hardware components are defined as services in DSS. During runtime, users execute synchronization in CCR. Therefore, the software and hardware of the robot is easy to modify or reuse for different applications. In comparison to previous works, the hardware and software platforms of this robot provide greater flexibility. The robot platform, originally purpose-built for pavement inspections, has now been generalized such that minimal modification would be required for use in different scenarios or applications - reusability of the design has become the central focus. Finally, practical experiments were performed to validate the efficacy of the motion modes. The satisfactory results indicated that integrating pavement inspection research and the autonomous robot has great potential. And we are hopeful the innovative method may be a basis for future researchers in the same field.

參考文獻


Abaza, K. A., Ashur, S. A., & Al-Khatib, I. A. (2004). Integrated Pavement Management System with a Markovian Prediction Model. Journal of Transportation Engineering, ASCE, Vol. 130, No. 1 , pp. 24-33.
[3] ASTM D5340. (2004). Standard Practice of Roads and Parking Lots Pavement Condition Index Surveys.
[4] ASTM D6433. (2003). Standard Practice of Roads and Parking Lots Pavement Condition Index Surveys.
[6] Diosi, A., & Kleeman, L. (2007, October 10). Fast Laser Scan Matching using Polar Coordinates. The International Journal of Robotics Research Vol. 26 , pp. 1125-1153.
[11] Liptak, B. (1995). Instrument Engineers 'Handbook: Process Control'. Radnor, Pennsylvania: Chilton Book Company, 20-29.

延伸閱讀