English  |  正體中文  |  简体中文  |  Items with full text/Total items : 26988/38789
Visitors : 2347184      Online Users : 29
RC Version 4.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Adv. Search
LoginUploadHelpAboutAdminister

Please use this identifier to cite or link to this item: http://ntour.ntou.edu.tw:8080/ir/handle/987654321/32106

Title: 運用加速規於載具定位
Using Accelerometer in Vehicle Positioning
Authors: 林鎮洲;楊智任;張文桐
Contributors: NTOU:Department of Mechanical and Mechatronic Engineering
國立臺灣海洋大學:機械與機電工程學系
Keywords: 卡爾曼濾波器;類神經網路;載具定位;加速規
Kalman filtering;Neural network;Vehicle positioning;Accelerometer
Date: 2007-11
Issue Date: 2012-06-15T07:23:26Z
Publisher: 中國機械工程學會第二十四屆全國學術研討會論文集
Abstract: 本論文以卡爾曼濾波器演算法為基礎,透過加速規對載具進行位置估測,並於研究中實際於線性運動系統上進行動態實驗。一般利用加速規進行載具定位時,由於感測器回授之信號含有雜訊,如直接將加速度經由二次積分計算出載具位置,則此位置估算值將會隨著時間發散。因此通常在進行積分之前必須先利用卡爾曼濾波器對加速度信號進行雜訊濾除,以獲得較準確之位置估測值。
  運用卡爾曼濾波器之前提為必須對狀態雜訊和量測雜訊之協方差矩陣形式能完全掌握,以獲得最佳之估測結果。故本論文透過設計實驗來量測此線性運動系統在各種急跳度下的最佳狀態雜訊值,並將此相對關係記憶於類神經網路中。之後於位置估測實驗時,卡爾曼濾波器便可以隨時依照目前的運動狀況更新狀態雜訊值。由實驗結果證實此演算架構可獲得不錯之估算效果。
  In this thesis, the problem of vehicle position estimation using an accelerometer is investigated based on Kalman filtering theory, and a linear motion system is implemented in the experiment to test the performance. In generally, the position estimation problems using accelerometers without any filtering process will become divergent as time elapse. This is because the sensor’s signals inevitably contain a certain extent of noises, and the noises accumulate during the integration process. Hence in order to estimate the position more accurately, Kalman filtering algorithm is usually employed to derive a more reliable result.
  One of the prerequisites in using Kalman filter is that the information of the process noise covariance and measurement noise covariance should be known in advance in order to derive the best estimation result. The above mentioned noises are generally dependent variables with respect to the motion states. Hence in the thesis, we measured the optimal state noises versus a set of jerk values through the experiments, and then by feeding them into the neural network to memorize the relationship. Then Kalman filter would update the state noise values according to the current motion states during the position estimation experiments. From the experimental result, it showed that the proposed algorithm was proved to be a reliable estimation algorithm.
Relation: 論文編號B14-0016, pp.1743-1749
URI: http://ntour.ntou.edu.tw/handle/987654321/32106
Appears in Collections:[機械與機電工程學系] 演講及研討會

Files in This Item:

File Description SizeFormat
index.html0KbHTML450View/Open


All items in NTOUR are protected by copyright, with all rights reserved.

 


著作權政策宣告: 本網站之內容為國立臺灣海洋大學所收錄之機構典藏,無償提供學術研究與公眾教育等公益性使用,請合理使用本網站之內容,以尊重著作權人之權益。
網站維護: 海大圖資處 圖書系統組
DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback