慣性和地磁感測陣列的無人機導航;Unmanned Aerial Vehicle Navigation with Inertial and Geomagnetic Sensor Array

NCU Institutional Repository > 資訊電機學院 > 資訊工程學系碩士在職專班 > 博碩士論文 > Item 987654321/95434

jsp.display-item.identifier=請使用永久網址來引用或連結此文件: http://ir.lib.ncu.edu.tw/handle/987654321/95434

题名:	慣性和地磁感測陣列的無人機導航;Unmanned Aerial Vehicle Navigation with Inertial and Geomagnetic Sensor Array
作者:	楊世鴻;Yang, Shih-Hong
贡献者:	資訊工程學系在職專班
关键词:	無人機;慣性導航;遞迴神經網路;UAV;Inertial navigation system;RNN
日期:	2024-07-23
上传时间:	2024-10-09 16:51:27 (UTC+8)
出版者:	國立中央大學
摘要:	無人機在無GNSS環境下，通常使用電腦視覺方法或是慣性導航。然而電腦視覺方法容易被環境光源影響。而現行的慣性導航有雜訊與誤差的問題，精度較高的慣性感測器有著成本高且體積大的缺點。因此我們採用MIAT系統設計方法論設計了高精度且低成本的無人機導航系統，它是基於一個慣性與地磁感測陣列及一個微控制器，透過校正演算法、卡爾曼濾波器、深度學習模型來降低導航的誤差。實驗結果顯示，我們的方法比起單一慣性感測器，在水平方向的矩形路徑實驗中平均定位誤差從11.3公尺降至1.65公尺，以及水平方向的圓形路徑實驗中平均定位誤差從7.11公尺減少至2.87公尺。此外在無人機實驗中，經過90秒的飛行後平均定位誤差為1.65公尺，證明我們的無人機導航系統的優秀性能。;In environments without GNSS, drones typically rely on either computer vision techniques or inertial navigation systems. However, computer vision can be significantly affected by environmental lighting, while current inertial navigation systems suffer from noise and inaccuracies. Higher precision sensors are also costly and bulky. To address these challenges, we designed a high-accuracy, low-cost drone navigation system using the MIAT system design methodology. Our system is based on an array of inertial and geomagnetic sensors, along with a microcontroller. It employs a calibration algorithm, filters, and a deep learning model to reduce navigational errors. Experimental results demonstrate that our method significantly reduces the average position error in a horizontal rectangular path experiment from 11.3 meters to 1.65 meters, and in a horizontal circular path experiment from 7.11 meters to 2.87 meters. Additionally, during drone experiments, after 90 seconds of flight, the average position error was 1.65 meters, proving the superior performance of our drone navigation system.
显示于类别:	[資訊工程學系碩士在職專班 ] 博碩士論文

文件中的档案:

档案	描述	大小	格式	浏览次数
index.html		0Kb	HTML	100	检视/开启

在NCUIR中所有的数据项都受到原著作权保护.

社群 sharing

数据加载中.....