Research and development of the map-guided automated vehicle
Date Issued
2015
Date
2015
Author(s)
Chao, Ko-Chin
Abstract
This research aims to develop a map-guidance technology for road vehicle automation using high accuracy digital maps. The objective is to utilize mid/low cost global navigation satellite system (GNSS), inertial navigation sensors and digital map data to achieve real-time precision positioning, navigation and motion control through data fusion techniques, rather than rely on costly laser sensors and the commonly adopted Simultaneous localization and mapping (SLAM) approach. The ultimate goal is to apply the proposed map-guidance technology to developing cost-effective, mid~low speed intelligent vehicles, such as unmanned shuttles on campus and the personal rapid transit (PRT) system. In order to achieve the abovementioned objectives, this thesis proposes a real-time vehicle position and altitude estimation algorithm based on the Constrained Unscented Kalman Filter (CUKF) technique, and the Check Point vehicle position error correction method, which both exploit digital map data to enhance the accuracy and reliability of the vehicle positioning and motion control systems. The proposed map-guidance vehicle automation technique has been implemented on an electric golf cart and tested on the campus of the National Taiwan University. The developed automated golf cart employs an economical single-frequency GNSS receiver, which can receives both U.S. GPS and China BeiDou satellites’ signals. Experimental results show that the lateral position estimation errors can be bounded within one meter, and the lane-level vehicle motion control precision can be achieved with the proposed methods.
Subjects
Data fusion
Map Guide
Kalman filter
Autonomous driving
Type
thesis
File(s)![Thumbnail Image]()
Loading...
Name
ntu-104-R01522825-1.pdf
Size
23.54 KB
Format
Adobe PDF
Checksum
(MD5):a4d504274964ee61129eb7aa21c7d087