Underwater map-based localization using flow features

详细信息    查看全文

• 作者：Naveed Muhammad ; Gert Toming ; Jeffrey A. Tuhtan ; Mark Musall…
• 关键词：Flow sensing ; Flow ; feature extraction ; Map ; based localization ; Underwater robotics
• 刊名：Autonomous Robots
• 出版年：2017
• 出版时间：February 2017
• 年：2017
• 卷：41
• 期：2
• 页码：417-436
• 全文大小：
• 刊物类别：Computer Science
• 刊物主题：Robotics and Automation; Artificial Intelligence (incl. Robotics); Computer Imaging, Vision, Pattern Recognition and Graphics; Control, Robotics, Mechatronics;
• 出版者：Springer US
• ISSN：1573-7527
• 卷排序：41

文摘

Underwater robots conventionally use vision and sonar sensors for autonomous localization. Fish on the other hand also have the ability to sense flow, which assists them in navigating. Recently, it has been shown that bioinspired flow sensing can be used in robotics, for tasks such as object detection and positioning in laboratory conditions. In this paper we present a map-based localization technique using flow sensing. The technique is based upon compact histogram features that are extracted from frequency spectra of pressure data acquired using a single piezo-resistive sensor. The features are used to create flow-based map of an underwater environment, and later during an off-line localization phase, similar features are again extracted and used inside a particle filter in order to perform localization. Experiments carried out using pressure data acquired inside a model fish pass validate the proposed technique.