An Efficient Safety Confirmation Method Using Image Database in Multiple-MDRU-Based Disaster Recovery Network

Safety confirmation is one of the most important applications of disaster-resilient networking based on the movable and deployable resource unit (MDRU). With an embedded image database, a single MDRU can provide the safety confirmation application that allows users to search for the images that look...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE systems journal 2017-12, Vol.11 (4), p.2556-2565
Hauptverfasser: Thuan Ngo, Nishiyama, Hiroki, Kato, Nei, Sakano, Toshikazu, Takahara, Atsushi
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Safety confirmation is one of the most important applications of disaster-resilient networking based on the movable and deployable resource unit (MDRU). With an embedded image database, a single MDRU can provide the safety confirmation application that allows users to search for the images that look similar to the people they are looking for. However, a network with multiple MDRUs may increase the image searching time of the safety confirmation application because the image databases are distributed and the capacity of the backbone wireless network constructed by connecting the MDRUs is also limited. Therefore, in this paper, we propose a safety confirmation method that guarantees the minimum searching time of users. The method includes four phases, namely, resizing and storing images, broadcasting small-size images, routing, and deciding image size to deliver to users. Based on mathematical analysis using the absorbing Markov chain, we estimate the expected searching time for each different size of images and choose the most appropriate size. Furthermore, we conduct extensive computer-based simulations to verify the findings of our analysis. The simulation results prove the existence of the optimal image size that minimizes the searching time and demonstrate the effectiveness of our proposed method.
ISSN:1932-8184
1937-9234
DOI:10.1109/JSYST.2015.2457777