Abstract at Time of Award

Disasters are events with dire consequences, requiring multiple-agency responses and resources beyond the capability of a single community. Natural disasters, such as the 2011 Great East Japan Earthquake, can threaten the lives of many people and cause inordinate economic losses. Communication is critical to disaster preparation, response, and recovery, but may be damaged during the disaster. In this project, researchers from the US and Japan study novel approaches to disaster preparation, response and recovery using survivable communication networks and big data analysis of social media data. This collaborative effort involves expertise in disaster research, social media mining and big data analysis, network science, wireless communications, and machine learning, to examine resilient network architecture and algorithms, data collection and analysis before the disaster, and decision making and information dissemination during the disaster. The resilient network incorporates both wired and wireless communications to deal with multiple disaster-induced failures, aiming for efficient algorithms serving emergency applications. State-of-the-art data collection and analysis techniques will help build an important knowledge base in proactive preparation for disasters. Real time decision making and information dissemination during a disaster can assist disaster response and recovery effectively.

The proposed research aims to provide valuable guidance for disaster preparation, response, and recovery for both the US and Japan, and spearhead a new research direction in survivable communication network design and big data analysis. This project provides a conducive environment to further research collaboration of big data analysis and disaster relief between the US and Japan. Graduate students will be jointly trained in this international research project to actively collaborate in carrying out the proposed research tasks. Special efforts will be made to engage minority students and underrepresented groups.

Summary of Research Outcome

Intellectual Merit:

The outcomes have been made available to research community through high quality journal articles and conference presentations. The research has resulted in five (5) journal papers, five (5) conference papers, and two (2) journal papers under second round of review. The ten published papers include two papers in the IEEE/ACM Transactions on Networking, one paper in IEEE Journal on Selected Areas in Communications, two papers in IEEE INFOCOM, and one paper in IEEE Globecom.

Natural disasters can result in severe damage to communication infrastructure, which leads to further chaos to the damaged area. After the disaster strikes, most of the victims would gather at the evacuation sites for food supplies and other necessities. Having a working communication network is very important to help the victims. We aim at recovering the network from the still-alive mobile base stations to the out-of-service evacuation sites by using multi-hop relaying technique. We formulated the population-aware relay node placement problem and studied various algorithms for solving this problem.

During natural or man-made disasters, people tweet warnings, request for help, and report about the environment. Millions of posts are published on social media every day during a disaster, reflecting on natural and man-made major events such as Hurricane Sandy. Among these posts, there are useful information such as live reports from people on site, requests for help, and offer for resources which are valuable for first responders. However, this data is usually overwhelmed with irrelevant information, spams, disinformation, misinformation, rumors, and bot-generated content. To gain accurate insights in the aftermath of a crisis, aforementioned harmful posts need to be removed. These include spams, misinformation, rumors, bot-generated content. We introduced these groups of posts and emphasized the importance of removing them along with brief overview of proposed methods to do so.

Broader Impacts:

The proposed research activities have complemented and enriched the growing curriculum on game theory and optimization at Arizona State University through course development and special topic seminars.

We have established strong research collaboration between the researchers in the USA and Japan. Both the PI and co-PI from the USA visited the collaboration institute in Tokyo (Japan). The PIs from Japan have visited ASU to work on the project. In addition, the PIs have met at IEEE conferences (such as IEEE Globecom2015 in San Diego, and IEEE Globecom2017 in Singapore). In addition, the two PIs from USA and Japan have organized a special session on Big Data and Disaster Management during DSAA2017 – The 4th IEEE International Conference on Data Science and Advanced Analytics. titled “Building Robust Wireless Networks with Relay Node Placement” at the 5^th IEEE International Conference on Information and Communication Technologies for Disaster Management, December 4-7, 2018, Sendai, Japan.

Highly skilled personnel in related areas have been trained in carrying out the proposed research tasks. Special efforts have been made to engage minority and underrepresented groups. Two Ph.D. students, Tahora Nazer and Ruozhou Yu have been involved in this project at Arizona State University. Ruozhou Yu will join North Carolina State University as a Tenure-Track Assistant Professor of Computer Science in Fall 2019.

Last Modified: 06/25/2019
Modified by: Guoliang Xue