蜜蜂搜尋扇形掃描混合雷達的目標偵測,定位及分群之資料處理

Data Processing with Target Detection, Localization, and Clustering for Sector Scanning Mixing Radar of Bee Searching

程懋川(Mao-Chuan Cheng)

指導教授 ： 連豊力

國立臺灣大學/電機資訊學院/電機工程學研究所/碩士(2020年)

https://doi.org/10.6342/NTU202000197

摘要

本篇論文為蜜蜂搜尋扇形掃描混合雷達提出一個完整資料處理的過程。完整的資料處理過程包含三個部分。首先，在目標偵測的部分，峰值可以用於確定來自混合雷達的相關數據是否是從標記目標反射的數據。峰值閥值在實驗中確定為0.1。在峰值閥值為0.1的情形下，距離的最大誤差在2公尺以內。如果來自混合雷達的相關數據的峰值大於峰值閥值，則相關數據被視為目標數據。然後，在目標定位的部分，目標數據用距離和方位角來定位。距離由相關數據的峰值位置去計算，方位角是讀取馬達的角度。因此，接收相關數據和讀取馬達角度的時間需要一致。提出了一個實驗來檢查峰值和方位角是否一致。最後，在目標分群的部分，每個目標數據包含距離、接收時間、方位角。根據扇形掃描混合雷達的特性，提出一種基於接受時間差和方位角的方法，將所有目標數據分類為兩種情況。一種情況是目標停留在固定位置，另一種情況是目標正在飛行。同時，箱形圖用於表示目標停留在固定位置的情況，單個點來用表示目標正在飛行的情況。同時本論文中提出了方法，解決了實驗中訊號微弱的問題。通過實際的蜜蜂實驗，本論文對所提出的資料處理過程進行了驗證。

關鍵字

資料處理 ； 目標偵測 ； 目標定位 ； 目標分群 ； 扇形掃描混合雷達 ； 蜜蜂搜尋 ； 箱形圖

並列摘要

In this thesis, a complete data processing is proposed for sector scanning mixing radar of bee searching. The complete data processing contains three parts. First of all, in the target detection, the peak value can be used to determine whether the correlation data from the mixing radar is the data reflected from the tagged target. The peak value threshold is determined by 0.1 in the experiment. Under the peak value threshold of 0.1, the maximum error of the distance is within two meters. If the peak value of the correlation from the mixing radar is more than 0.1, the correlation is regarded as the target data.
Then in the target localization, the target data is located by the distance and the azimuth. The distance is calculated by the peak location of the correlation data. The azimuth is determined by reading the motor angle. Therefore, the time of receiving the correlation data and reading the motor angle need to be consistent. An experiment is proposed to check whether the peak value and the azimuth are consistent.
Finally, in the target clustering, each target data contains the distance, the azimuth, and the received time. According to the property of sector scanning mixing radar, a method based on the received time difference and the azimuth is proposed to classify all target data into two situations. One situation is that the target is staying in a fixed position, and the other situation is that the target is flying. At the same time, a boxplot is used to show the situation that the target is staying in a fixed position and a single point is used to show the situation that the target is flying. In addition, a method is proposed to overcome the problem of the weak signal in the experiment in this thesis.
Bee Experiments are provided to verify the proposed data processing process.

並列關鍵字

Data processing ； target detection ； target localization ； target clustering ； sector scanning mixing radar ； bee searching ； boxplot

參考文獻

[1: Tsai et al. 2013]

Zuo-Min Tsai, Pei-Hung Jau, Nai-Chung Kuo, Jui-Chi Kao, Kun-You Lin, Fan-Ren Chang, En-Cheng Yang, and Huei Wang, “A High-Range-Accuracy and High-Sensitivity Harmonic Radar Using Pulse Pseudorandom Code for Bee Searching,” IEEE Transactions on Microwave Theory and Techniques, Vol. 61, No. 1, pp. 666–675, Jan. 2013.

[2: Riley & Smith 2002]

J. R. Riley and A. D. Smith, “Design considerations for an harmonic radar to investigate the flight of insects at low altitude,” Computers and Electronics in Agriculture, Vol. 35, pp. 151–169, 2002.

[3: Psychoudakis et al. 2008]