調頻連續波雷達系統之低複雜度數位波束成型及靜態物辨識

Low Complexity Digital Beamforming and Static Object Identification for FMCW Radar System

林冠亨(Kuan-Hen Lin)

指導教授 ： 李大嵩

國立交通大學/電機學院/電信工程系所/碩士(2019年)

摘要

自駕車已成為未來汽車工業發展的趨勢，透過光達、毫米波雷達、影像感測器等各式感測技術可強化車輛對環境的感知能力與緊急事件的偵測與應變能力。由於使用環境限制及成本的考量，毫米波雷達偵測技術逐漸受到重視。為掌握目標物的資訊，目標物偵測的設計成為系統重要的一環。傳統的設計主要採用恆虛警率的偵測器，在偵測到目標物的距離與速度資訊後，再搭配角度估計器得到目標物的精準位置資訊。然而，在複雜的真實環境中，偵測效能除了會因為雜訊估計錯誤而下降，也會因為目標回波訊號能量較小而下降。在車用雷達的應用中，靜態物觸發的警示也被視為誤警率的一部分。在本論文中，吾人引入空間濾波器之技術並提出偵測程序以提升偵測效能並且降低偵測複雜度。在靜態物造成誤警問題中，吾人根據都普勒特性設計投影方式以辨識靜態物。經由模擬驗證，吾人所提出的偵測程序相較於傳統目標偵測系統在嚴苛環境下達到了較高的偵測品質及低複雜度。吾人提出的靜態物辨識方法也有相當程度的辨識率。

關鍵字

調頻連續波 ； 車用雷達 ； 低複雜度偵測演算法 ； 靜態物辨識 ； 數位波束成型

並列摘要

Self-driving has become the trend of the development in the automotive industry. Through LiDAR, mmWave radar and image sensors, these various sensing technologies enhance vehicle's perception of the environment. Due to the considerations of various scenarios and cost, increasing importance has been attached to mmWave radar. In order to grasp the information of the target precisely, the detector plays an important role. In conventional ones, the constant false alarm rate (CFAR) detector is adopted to acquire targets’ range and velocity information, then the angle estimator is used to obtain accurate position of targets. However, detection loss happens not only due to the error estimation of the noise, but also weak signal energy of the target echo. In addition, the static objects need to be identified, which are also considered as parts of the false alarm rate in automotive radar applications. In this thesis, we propose a low complexity detection procedure which includes spatial filter to improve detection performance. In the false alarm problem caused by static objects, we design a projection method according to the Doppler characteristics to identify static objects. The simulation results show that the proposed detection procedure elevates the detection performance in the harsh environment, while reducing the computational complexity. Moreover, the simulation results also show the proposed static object identification method is reliable.

並列關鍵字

CFAR detector ； Digital beamforming ； FMCW radar ； MTI ； Static object identification

參考文獻

[1] B. Sinzig, “Forward Collision Accidents: The (Swiss) Insurance Company Perspective,” AXA Winterthur, Feb., 2009

[2] J. Wenger, “Automotive Radar—Status and Perspectives,” in Proc. of IEEE CSIC, 2005, pp. 21–24.

[3] D. Borrick, “FM/CW Radar Signal and Digital Processing,” NOAA Technical Reports ERL283-WPL20, Jul. 1973.

[4] A. Stove, “Linear FMCW Radar Techniques,” IEEE Proc. F, vol. 139, pp. 343-350, Oct. 1992.

[5] S. Miyahara, “New Algorithm for Multiple Object Detection in FM-CW Radar,” SAE Technical paper series, Mar. 2004.