簡易檢索 / 詳目顯示

回結果列表
研究生： 林宸玄
Lin, Chen-Xuan
論文名稱： 基於資料流程適用於 ROS 自動駕駛軟體之效能分析工具
Data Flow Aware Profiling for ROS­-based Autonomous Vehicle Software
指導教授： 涂嘉恒
Tu, Chia­-Heng
學位類別： 碩士
Master
系所名稱： 電機資訊學院 - 資訊工程學系
Department of Computer Science and Information Engineering
論文出版年： 2021
畢業學年度： 109
語文別： 英文
論文頁數： 29
中文關鍵詞： Robot Operating System 2 、自動駕駛軟體 、效能指標 、資料流程 、效能分析工具
外文關鍵詞： Robot Operating System 2, autonomous software, self-­driving, performance tool, performance characterization
相關次數： 點閱：188  下載：0 
分享至:
查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報

    • 隨著自動駕駛技術的逐漸發展，基於 Robot Operating System 2 (ROS2) 的自動駕駛軟體的複雜性也在增加。對於系統設計師來說，快速了解此類複雜軟體的運作行為和效能是一個巨大的挑戰，雖然傳統工具可以顯示系統的效能狀態，但不足以表達這種模組化且複雜軟體中潛在的效能問題。在本論文中透過資料流程分析了兩種效能問題來表達 ROS2 模組裡是否存在潛在的效能問題, 以及顯示相關的效能統計數據。我們將資料流程的效能分析應用在自動駕駛軟體 ──Autoware.auto，得到的結果令人鼓舞。

    The complexity of the Robot Operating System 2 (ROS2) based autonomous software grows as autonomous vehicles technology has gradually developed. It is a big challenge for system designers to rapidly understand the runtime behaviors and performance of such sophisticated software because the conventional tools are insufficient for characterizing potential performance issues of the modules within the software. In this thesis, two performance issue analyses are devised to represent the potential peformance issues and related performance statistics of the ROS modules. The data flow aware profiling is applied to the autonomous software Autoware.auto, and the results are encouraging.

    摘要 i Abstract ii 誌謝 iii Table of Contents iv List of Tables v List of Figures vi Chapter 1. Introduction 1 Chapter 2. Background and Motivation 6 2.1 Robot Operating System 6 2.2 ROS-­based Autonomous Driving Software 8 2.3 Extended Berkeley Packet Filter 9 Chapter 3. Methodology 13 3.1 ROS and non­-ROS Clasification 14 3.2 Data Flow Aware Profiling 14 3.2.1 CPU Bound Analysis 17 3.2.2 Data Frequency Bound Analysis 18 Chapter 4. Experimental Results 21 4.1 Experimental Setup 21 4.2 CPU­-bound analysis 21 4.3 Data frequency bound analysis 25 Chapter 5. Conclusion 28 References 29

    [1] eBPF. Introduction, Tutorials, Community Resources, July 2020.
    [2] EfficiOS. LTTng: Tracing framework for Linux., July 2020.
    [3] Autoware foundation. Autoware.Auto, July 2020.
    [4] Open Source Robotics Foundation. Project Governance, July 2020.
    [5] Kernel.org. perf: Linux profiling with performance counters, October 2019.
    [6] LG Electronics. LGSVL Simulator, October 2019.
    [7] Ingo Lütkebohle and Christophe Bedard. ros2tracing: Tracing tools for ROS 2., July 2020.
    [8] OMG. Data Distribution Service, July 2020.
    [9] IO Visor Project. toolkit for enabling efficient kernel tracing, July 2020.
    [10] Morgan Quigley, Brian Gerkey, Ken Conley, Josh Faust, Tully Foote, Jeremy Leibs, Eric Berger, Rob Wheeler, and Andrew Ng. Ros: an open­source robot operating system. In ICRA, page 5, May 2009.
    [11] Dirk Thomas. Ros wiki: rqt_graph, October 2019.
    [12] Lijun Wei, Cindy Cappelle, and Yassine Ruichek. Camera/laser/gps fusion method for vehicle positioning under extended nis­based sensor validation. 62:3110–3122, November 2013.

    下載圖示
    2026-07-19公開
    QR CODE