跳到主要內容

臺灣博碩士論文加值系統

(44.200.86.95) 您好!臺灣時間:2024/05/30 06:16
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果 :::

詳目顯示

我願授權國圖
: 
twitterline
研究生:李俊穎
研究生(外文):LEE, JUN-YING
論文名稱:無人機視覺之即時路徑重劃及障礙物地理資訊建立
論文名稱(外文):Visual Drone Real-Time Rerouting and Obstacle Geographic Information Creation
指導教授:李孟澤李孟澤引用關係
指導教授(外文):LEE, MENG-TSE
口試委員:賴盈誌呂文祺李孟澤
口試委員(外文):LAI, YING-CHIHLU, WEN-CHILEE, MENG-TSE
口試日期:2022-07-28
學位類別:碩士
校院名稱:國立虎尾科技大學
系所名稱:自動化工程系碩士班
學門:工程學門
學類:電資工程學類
論文種類:學術論文
論文出版年:2022
畢業學年度:110
語文別:英文
論文頁數:55
中文關鍵詞:邊緣運算立體視覺路徑重劃禁航區
外文關鍵詞:edge computingstereo visionpath re-planningno-fly zone
相關次數:
  • 被引用被引用:0
  • 點閱點閱:109
  • 評分評分:
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
本研究主要對障礙物避讓的部分進行優化,使無人機在進行避讓時,同時間將深度相機所擷取的未知障礙物訊息以障礙物區塊的形式記錄下來,並透過Wi-Fi將資訊上傳至雲端資料庫,最後透過障礙物區塊拼接演算法(obstacle block stitching algorithm)對障礙物訊息進行整理並優化,逐漸完整化障礙物的資訊。
障礙物區塊資訊的處理是透過無人機上的Xavier NX嵌入式電腦,根據所辨識之未知障礙物進行區塊的推算並將其定義為障礙物區塊,接著透過雲端資料庫中的障礙物區塊拼接演算法對障礙物區塊做處理,當有障礙物區塊發生重疊或相鄰時,將依據條件進行障礙物區塊的拼接及整合。使無人機在執行相同區域的飛行任務時,將可藉由雲端資料庫獲取此區域的障礙物資訊,並透過路徑規劃演算法規劃任務路徑,防止無人機對同一障礙物進行再次避讓,以提升無人機的飛行效率及飛航安全。
在最後的驗證實驗中,我們建立了一架具有邊緣運算能力的無人機作為驗證載具。實驗中我們對障礙物進行多方位的飛行以模擬同區域下不同的飛行任務,並藉此進行雲端資料庫以及障礙物區塊拼接演算法的驗證,最後也成功地驗證了這套系統的功能。

This study focuses on the optimization of the obstacle avoidance program, so that when the UAV is avoiding the unknown obstacle, its information is captured by the depth camera and recorded in the form of obstacle block. Then the information is uploaded to the cloud database via Wi-Fi, where the obstacle information is collated and optimized by the obstacle block stitching algorithm to complete the obstacle information in detail.
The obstacle block information is processed by the Xavier NX embedded computer equipped on the UAV, which calculates the block and defines it as an obstacle block according to the recognized unknown obstacle. Then, the information of the obstacle block is processed by the obstacle block stitching algorithm in the cloud database. When there are obstacle blocks overlapping or adjacent to each other, the obstacle blocks will be stitched and integrated according to the conditions. This allows the UAV to obtain information about the obstacles in a zone through the cloud database when performing a mission in the same area, and to plan the mission path through the path planning algorithm to prevent the UAV from avoiding the same obstacle again, thus enhancing the usage rate of the UAV.
In the final validation experiment, a UAV with edge computing capability was built as a test vehicle. In the experiment, we flew multiple directions of the obstacle to simulate the path of different tasks in the same area. This is used to validate the cloud database and the obstacle block stitching algorithm. In the end, the system was successfully validated.

摘要......i
Abstract......ii
Acknowledgements......iv
Table of Contents......v
List of Tables......vii
List of Figures......viii
Symbols......x
Chapter 1 Introduction......1
1.1 Background Information......1
1.2 Motivation......2
1.3 Research Purpose......3
1.4 Research Framework......4
Chapter 2 Literature Survey......5
2.1 Obstacle Detection......5
2.2 Real-Time Path Planning Algorithm......6
2.3 No-Fly Zone Avoidance Algorithm......6
Chapter 3 Research Content and Methods......8
3.1 UAV System Architecture......8
3.2 Image Recognition......9
3.2.1 Fundamentals of Stereo Vision......9
3.2.2 Obstacle Recognition......11
3.3 Obstacle Avoidance......16
3.3.1 3D Obstacle Block Establishment and Coordinates Calculation......16
3.3.2 3D Obstacle Block Establishment and Coordinates Calculation Steps......18
3.3.3 Real-Time Rerouting Algorithm......19
3.4 Obstacle Block Stitching Algorithm......20
3.4.1 Obstacle Block Stitching Process......21
3.4.2 Obstacle Block Stitching Algorithm Steps......22
3.5 Cloud Database......23
3.5.1 Data Collection......23
3.5.2 Calculated Obstacle Blocks Coordinates......24
3.5.3 Stitched Obstacle Block Coordinates......24
Chapter 4 Hardware Structure and Integration......26
4.1 Test Drone......26
4.2 Hardware Specifications......27
4.2.1 Pixhawk4 Flight Controllers......27
4.2.2 Radio Control System......28
4.2.3 Digi XBee PRO Wireless Communication Module......28
4.2.4 Components of Drone......29
4.2.5 Jetson Xavier NX Edge Computing System......30
4.2.6 Intel RealSense Depth Camera D455......31
Chapter 5 Experiment Planning and Results......32
5.1 Experimental Plan......33
5.2 Experiment Verification......33
5.2.1 Avoidance Experiment......33
5.2.2 Obstacle Blocks Stitching Experiment......40
Chapter 6 Conclusions and Future Study......47
6.1 Conclusion......47
6.2 Future Study......47
Reference......49
Extended Abstract......51
[1]The register, A Wing drone image
https://www.theregister.com/2022/04/04/texas_wing_drone/
[2]Bo-Yu Chen, “Failure Robot Path Complementation for Robot Swarm Mission Planning”, Master Thesis, Automation Engineering, National Formosa University, Yunlin, 2017.
[3]Wun-Siang Huang, “Task Assignment of UAV Swarm with No-fly Zone Avoidance”, Master Thesis, Automation Engineering, National Formosa University, Yunlin, 2020.
[4]Star network, Retrieved May 11, 2022, from
https://en.wikipedia.org/wiki/Star_network
[5]Yan-Ru Chen, “UAV Swarm Mission Planning and Real-Time Rerouting Under Uncertain Environment”, Master Thesis, Automation Engineering, National Formosa University, Yunlin, 2021.
[6]OpenCV, Retrieved May 11, 2022, from
https://opencv.org/
[7]R. Girshick, J. Donahue, T. Darrell and J. Malik, “Rich feature hierarchies for accurate object detection and semantic segmentation”, University of California, Berkeley, USA, 2014.
[8]J. Redmon, S. Divvala, R. Girshick, A. Farhadi, “You Only Look Once: Unified, Real-Time Object Detection”, University of Washington, Allen Institute for AI, USA, 2015.
[9]J. Borenstein and Y. Koren, “The vector field histogram-fast obstacle avoidance for mobile robots”, IEEE Transactions on Robotics and Automation, vol. 7, no. 3, pp.278-288, 1991. doi: 10.1109/70.88137
[10]I. Ulrich and J. Borenstein, “VFH+: Reliable Obstacle Avoidance for Fast Mobile Robots”, The University of Michigan Advanced Technology Laboratories, Belgium, 1998.
[11]I. Ulrich and J. Borenstein, “VFH *: Local Obstacle Avoidance with Look-Ahead Verification”, The University of Michigan, USA, 2000.
[12]S. Koenig, M. Likhachev, “D* Lite”, College of Computing, Georgia Institute of Technology, Atlanta, School of Computer Science, Carnegie Mellon University, Pittsburgh, 2002.
[13]Pixhawk4, Image and Datasheet, Retrieved Jan 04, 2022, from
https://docs.px4.io/master/en/flight_controller/pixhawk4.html
[14]Jetson Xavier NX Developer Kit. Datasheet, Retrieved May 12, 2022, from
https://developer.nvidia.com/embedded/jetson-xavier-nx-devkit
[15]Intel RealSense Depth Camera D455. Datasheet, Retrieved May 12, 2022, from
https://www.intelrealsense.com/depth-camera-d455/
[16]Basic principles of depth camera. Retrieved Jan 08, 2022, from
https://blog.csdn.net/Yong_Qi2015/article/details/107527366
[17]A. Stentz, “The Focussed D* Algorithm for Real-Time Replanning”, Robotics Institute, Carnegie Mellon University, USA,1995

電子全文 電子全文(網際網路公開日期:20270823)
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top