Deep Learning Drone Flying Height Prediction for Efficient Fine Dust Concentration Measurement

被引：3

作者：

Yoon, Ji Hyun ^{[1
]}

Li, Yunjie ^{[1
]}

Lee, Moon Suk ^{[1
]}

Jo, Minho ^{[2
]}

机构：

[1] Korea Univ, Dept Comp & Informat Sci, Sejong Metropolitan City, South Korea

[2] Korea Univ, Dept Comp Convergence Software, Sejong Metropolitan City, South Korea

来源：

PROCEEDINGS OF THE 13TH INTERNATIONAL CONFERENCE ON UBIQUITOUS INFORMATION MANAGEMENT AND COMMUNICATION (IMCOM) 2019 | 2019年 / 935卷

关键词：

Fine dust concentration; Drone; Deep learning; RNN; CNN; Prediction of maximum flying height;

D O I：

10.1007/978-3-030-19063-7_88

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

Fine dust concentration has been collected so far by fine dust concentration monitoring towers at fixed heights. However fine dust concentration level varies significantly with heights. It is possible for people to get informed of wrong dust concentration information. Drone equipped with a fine dust sensor can fly up and down to sense fine dust concentration. Drone can solve the wrong fine dust concentration information problem. But we face too much energy consumption problem of drone and possibly delayed information because drone should fly from ground up to top. To solve this problem, we propose to cut drone flying height by predicting the height, using deep learning methods, at which maximum fine dust concentration can be sensed. Experimental results show that the proposed methods save 58.28% of flying distance.

引用

页码：1112 / 1119

页数：8