由於近年來電動車、自駕車的科技技術演進，利用車用感測器快速探測障礙 物成為近期熱門討論的話題。車用感測器對於車用輔助系統例如主動式巡航定速 系統(Adaptive Cruise Control, ACC)、自動緊急煞車系統(Autonomous Emergency Braking, AEB)、自動停車系統(Automatic Parking System, APS)等有著重大的幫 助。
在本篇論文中，將會先介紹聚類演算法，從中挑選兩種類別(劃分法及密度 法)後再選出兩種方法:一、劃分法_K-平均演算法(K-means)，二、密度法_ 基於密度的聚類演算法(Density-based spatial clustering of applications with noise, DBSCAN)。將兩種方法分別的介紹及說明其運作流程，接著完成初步模擬的結 果圖，最後再來判斷是否適合用於車用電子聚類演算法。
最終，選擇了 DBSCAN 來完成聚類演算法用於地圖繪製技術的可行性研究， 透過 C 程式語言以及 OpenGL 開放式圖型庫，成功模擬出障礙物位置的聚類結 果，以及完成區分雜訊的效果。證實 DBSCAN 演算法適合當作車用電子聚類演 算法的其中一種。

Due to the technological evolution of electric vehicles and self-driving cars in recent years, the use of to instantly detect obstacles has become a hot topic of discussion nowadays. Vehicle sensors are in a great help of Advanced Driver Assistance System (ADAS); for instance, Adaptive Cruise Control (ACC), Autonomous Emergency Braking (AEB), Automatic Parking System (APS), etc.
In this study, the clustering algorithm will be introduced first, from which two categories will be selected: division method and density method; then, two methods will be selected from those two categories: 1. K-means algorithm (K-means) 2. Density- based clustering algorithm (Density-based spatial clustering of applications with noise, DBSCAN). From the following paragraph, first of all, these two methods will be introduced and explained respectively. Then, the result diagram of the preliminary simulation will be completed. Last, the result diagram of K-means and DBSCAN will be analyzed whether it is suitable for the vehicle electronic clustering algorithm.
In the end, DBSCAN was selected to complete the feasibility study on using clustering algorithm for mapping technology. Through the C Programming Language and the OpenGL (Open Graphics Library), the clustering results of the obstacle positions were successfully simulated, and the method of distinguishing the noise was completed. It is confirmed that the DBSCAN algorithm is suitable as one of the vehicle electronic clustering algorithms.

[1] Sohee Lim, Seongwook Lee and Seong-Cheol Kim, "Clustering of Detected Targets Using DBSCAN in Automotive Radar Systems", 19 th International Radar Symposium (IRS 2018) , June 2018.
[2] Masanori Kunita, "Range Measurement in Ultrasound FMCW System", Electronics and Communications in Japan, Part 3, Vol. 90, No. 1, 2007
[3] Martin Ester, Hans-Peter Kriegel, Jörg Sander and Xiaowei Xu, "A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise", Int. Conf on Knowledge Discovery and Data Mining (KDD) Oregon, Aug 1996.
[4] J. Macqueen, "Some methods for classification and analysis of multivariate observations", 5th Berkeley Symp. Math. Statist. Prob, pp. 281-297, 1967.
[5] 阿新. (2019, February 4). 機器學習--聚類分析(劃分方法，層次方法、密度 方法). 程式人生. https://www.796t.com/content/1549253011.html
[6] 聚類分析. (2015, September 25). MBA 智庫百科. https://wiki.mbalib.com/zh- tw/聚类分析
[7] 林倢愷. (2021, April 7). 不要再用 K-Means! 超實用分群法 DBSCAN 詳解. Medium. https://axk51013.medium.com/不要再用 K-means-超實用分群法 dbscan 詳解-a33fa287c0e [8]Dbscan.(n.d.).MathWorks.https://www.mathworks.com/help/stats/dbscan.html#Ref erences
[9] 林嘉慶主持(2021)。測距感測器物體辨識與地圖繪製技術(同致電子企業股 份有限公司)。桃園，國立中央大學通訊工程學系。