╨╧рб▒с>■ ■ ■ Є╠с░┴т\pJava Excel API v2.6 Bфa=Ьп╝=h\:╛#8X@Н"╖┌1╚ РArial1╚ РArial1╚ РArial1╚ РArialрї └ рї └ рї └ рї └ рї └ рї └ рї └ рї └ рї └ рї └ рї └ рї └ рї └ рї └ рї └ р └ р+ї └ р)ї └ р,ї └ р*ї └ р ї └ р УА УА УА УА УА УА `ЕтDCМ,№Ftitle[*]contributor[author]contributor[advisor]keywords[*]date[issued] publisher citationsidentifier[uri]identifier[doi]abstractrelation[journal]relation[volume]relation[no]relation[page]ЫProbabilistic Smoothing based Generation of a Reliable Lane Geometry Map with Uncertainty of Lane Detectors;
(╠ ┴ АмЬ═0оX╟ И╜U╓ф┬1┴ рм$╕|╣ ╟\╒ U╓`╣╚ ╔╙─╨T╓ 0о╝ (╠ ┴ ╓┴└ └╔─│ ▌└1┴ )╝Х╝;Seokwon Kim ┴░╞Е║8╓2019-02\╒С┼│Y╒Pнxhttps://repository.hanyang.ac.kr/handle/20.500.11754/100390;
http://hanyang.dcollection.net/common/orgView/200000434687;This paper describes a generation method for a reliable lane geometry map. When a lane geometry map is generated, the lane curve, which is detected from a lane detector, is useful information because the lane geometry information can be obtained directly. However, the detected lane curve contains an uncertainty caused by the noise of the lane detector. In order to generate a reliable lane geometry map by accumulating the overall lane curve and considering the uncertainty of the detected lane curve, the probabilistic lane smoothing-based generation method for a reliable lane geometry map is proposed.
The proposed map generation method consists of four steps: data acquisition, vehicle trajectory estimation, probabilistic lane smoothing, and lane geometry modeling. Raw data including Global Positioning Systems (GPS) information, vehicle motion information, and the detected lane curve are acquired from the probe vehicle. The GPS information is converged with the vehicle motion information for improving the reliability of the GPS information. In the probabilistic lane smoothing step, the accumulated lane curves are smoothed based on a Bayesian filtering scheme. For smoothing the accumulated lane curves, the uncertainty of the lane curves is obtained from the sensor error model, and the lane geometry map is modeled as the nodes with the uncertainty of the position. The generated lane geometry map from the probabilistic lane smoothing step is represented as a B-spline curve for reducing the number of representations in the lane geometry modeling step.
The generated lane geometry map from the proposed method is verified and evaluated on a real road condition. The evaluation results show that the lane geometry map can be generated by reducing the noise of the detected lane curve. Additionally, the lane geometry map generated from the proposed method is more reliable in terms of the accuracy of the distance and the heading angle than the lane geometry map generated from the previous studies that use sampling point-based map generation methods. ▓ъ¤&HQsjМЕзЪ╝п╤╨Єё&)KLngЙЖи┐с╪·уцў&ў
Є╠Юf
d№йё╥MbP?_*+ВА% Б┴ГДб" ,,р?р?U ¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
¤
╫░─>╢@
■ ■ ■ ¤ ■ Root Entry ■ Workbook SummaryInformation( DocumentSummaryInformation8