利用光達及數位影像用於電動車避障路徑規劃

Obstacle avoidance path planning for electric vehicles using lidar and digital image

李承哲(Li Cheng-Che)

指導教授 ： 楊榮華

國立屏東科技大學/工學院/車輛工程系所/碩士(2022年)

https://doi.org/10.6346/NPUST202200203

摘要

本論文提出了一種利用光達和數位影像進行自動駕駛車輛避障的方法。該方法分為三個步驟，即障礙物檢測、目標追蹤與避障路徑規劃。光達和數位影像作為障礙物檢測的依據，利用點雲聚類與傳感器融合達成障礙物檢測之目的。並使用卡爾曼濾波器和匈牙利算法對車輛周圍的物體實現目標跟踪。而避障路徑規劃則是使用動態窗口法來規劃適當的避障路徑。
實驗平台為一輛由NVIDIA Jetson TX2嵌入式運算裝置作為控制器的小型電動車，並搭載GPS、六軸IMU、光達與攝像頭，進行定位、計算車輛偏航角與偵測車輛周遭環境資訊。實驗過程中，在低速、正確檢測障礙物的情況下，可以順利避開障礙物。

關鍵字

物件偵測 ； 傳感器融合 ； 物件追蹤 ； 避障路徑規劃

並列摘要

In this thesis, an obstacle avoidance method by using Lidar information and camera digital images for autonomous vehicle are proposed. The method is composed of three parts, namely, obstacle detection, object tracking, and path planning for obstacle avoidance. In obstacle detection part, point cloud clustering and sensor fusion are used to detect obstacles, while the object tracking part using Kalman filter and Hungarian algorithm to activate target tracking surrounding the vehicle. In the path planning for obstacle avoidance stage, the Dynamic Window Approach is utilized to program the most appropriate path for obstacle avoidance.
The experimental apparatus is an electric vehicle with NVIDIA Jetson TX2 embedded computing device as the controller. It is equipped with global positioning system module, six-axis inertial measurement unit, Lidar and camera. Experiments show the capability of avoiding obstacles smoothly under the conditions of low speed and correct obstacle detection.

並列關鍵字

Object detection ； Sensor fusion ； Object tracking ； Avoidance path planning

參考文獻

[1] John P.Snyder, 1987, Map Projections-A Working Manual, U.S. Government Printing Office, Washington, D.C.

[2] 趙伯勗，2018，運用點雲分割之多目標移動物體偵測與追蹤方法，碩士論文，國立雲林科技大學，資訊工程系，雲林。

[3] 伍育紅，2015，「聚類算法綜述」，計算機科學，第42卷，第6A期

[4] Daniel Eriksson, Jonas Harström, 2019, Object detection by cluster analysis on 3D-points from a LiDAR sensor, Master thesis, Department of Electrical Engineering, Chalmers University of Technology, Gothenburg, Sweden.

[5] Martin Ester, Hans-Peter Kriegel, Jiirg Sander, Xiaowei Xu, 1996, “A density-based algorithm for discovering clusters in large spatial databases with noise,” In Proceedings of the Second International Conference on Knowledge Discovery and Data Mining, AAAI Press, pp. 226–231.