This Rapid Response Research (RAPID) project is developing technology for ubiquitous event reporting and data gathering on the 2010 oil spill in the Gulf of Mexico and its ecological impacts. Traditional applications for monitoring disasters have relied on specialized, tightly-coupled, and expensive hardware and software platforms to capture, aggregate, and disseminate information on affected areas. We lack science and technology for rapid and dependable integration of computing and communication technology into natural and engineered physical systems, cyber-physical systems (CPS). The tragic Gulf oil spill of 2010 presents both a compelling need to fill this gap in research and a critical opportunity to help in relief efforts by deploying cutting-edge CPS research in the field. In particular, this CPS research is developing a cloud-supported mobile CPS application enabling community members to contribute as citizen scientists through sensor deployments and direct recording of events and ecological impacts of the Gulf oil spill, such as fish and bird kills.
The project exploits the availability of smartphones (with sophisticated sensor packages, high-level programming APIs, and multiple network connectivity options) and cloud computing infrastructures that enable collecting and aggregating data from mobile applications. The goal is to develop a scientific basis for managing the quality-of-service (QoS), user coordination, sensor data dissemination, and validation issues that arise in mobile CPS disaster monitoring applications.
The research will have many important broader impacts related to the Gulf oil spill disaster relief efforts, including providing help for the affected Gulf communities as they field and evaluate next-generation CPS research and build a sustained capability for capturing large snapshots of the ecological impact of the Gulf oil spill. The resulting environmental data will have lasting value for evaluating the consequences of the spill in multiple research fields, but especially in Marine Biology. The project is collaborating with Gulf area K-12 schools to integrate disaster and ecology monitoring activities into their curricula. The technologies developed (resource optimization techniques, data reporting protocol trade-off analysis, and empirical evaluation of social network coordination strategies for an open data environment) will provide a resource for the CPS research community. It is expected that project results will enable future efforts to create and validate CPS disaster response systems that can scale to hundreds of thousands of users and operate effectively in life-critical situations with scarce network and computing resources.
General Overview: The natural dispersion of humans, coupled with the prevalence of smartphone platforms, has created a great opportunity for data collection from the field. The success of ‘citizen scientist’ smartphone applications, which utilize smartphones to engage everyday participants in the process of collecting field data or survey results, has generated a desire for more smartphone applications intended for collecting data. This interest spans well beyond the computer science community-- individuals in healthcare, agriculture, disaster recovery, and other domains are looking to computer scientists and researchers for help in creating these applications. A key challenge of developing smartphone-based data collection citizens, however, is the high development cost, which limits their widespread use. For example, building data collection clients, cloud data collection infrastructure, privacy preserving mechanisms, collection protocol optimizations, and security policies is challenging. Moreover, the significant time required to develop these aspects of a smartphone data collection system make rapid development of applications highly customized to a specific disaster challenging. Finally, because of smartphone data collection systems’ development complexity they cannot be directly designed and created by scientists or first responders. The CLoud Environmental Analysis and Relief (CLEAR) system is a cross-platform configurable data collection system intended to allow rapid development of smartphone data collection applications without requiring a need for computer expertise. In any situation that could benefit from field data, such as a researcher interested in tracking the spread of a crop disease or an ecological disaster where field data would assist in understanding and mitigating ecological impact, CLEAR can be used by domain experts to rapidly create a data collection plan that outlines key information of interest, and deploy that plan to any smartphone participants in the affected area. The CLEAR web portal uses an intuitive building-block approach to defining data collection plans, allowing a plan to be composed of blocks such as ‘Acquire Location’, ‘Multiple Choice’, ‘Capture Picture’, and ‘Free Text Entry’, each of which comes with multiple configuration parameters. The CLEAR system can display these data entry blocks on smartphone devices, accept user input, optimize data reporting, aggregate and visualize results in a cloud, and coordinate with the CLEAR server to report data. Initially, CLEAR was developed and applied to the Gulf Oil Spill disaster context of 2010. After the oil spill clean-up was completed, a new context emerged for data collection needs by tornado damage assessment teams. The latter part of this project focused on developed a smartphone collection system for tornado assessment. Intellectual Merit: The main them of the CLEAR project was cloud and smartphone architectural techniques for rapidly developing sensor data collection applications for smartphones and pushing them to devices using a cloud infrastructure. The research work in CLEAR was performed in collaboration with by Dr. Jules White (at Virginia Tech during the project duration, and now at Vanderbilt University) and Dr. Aniruddha Gokhale (Vanderbilt). The project focused on 6 key areas of enabling citizen scientists and first responders to collect disaster data with smartphones: Modeling and simulation of smartphone data collection systems to verify and validate key data collection and power consumption QoS parameters Smartphone and cloud architectures for rapidly building and disseminating disaster data collection applications Improved K-anonymity algorithms for ensuring the privacy of citizen scientists without rendering data readings to imprecise Disaster data dissemination and mirroring protocol optimization for power efficiency in smartphone data collection systems Secure and simple sensor-based methods of ad-hoc data exchange between first responders and citizen scientists Algorithms for enabling citizen scientists with motor disabilities or first responders wearing gloves to interact with touchscreen sensor data applications using alternate accelerometer-based input mechanisms. The early years of the project produced specific contributions in the area of disaster management for oil spills, with the latter year concentrated on an adaptation focused on tornado assessment damage collection Broader Impacts: The project developed a prototype that showed the potential for involving the general citizenry to participate in the data collection process for different types of disasters. The project provided an opportunity to work with Civil Engineers who served as collaborators driving the needs for tornado damage tracking. High School science fair project focused on biology and ecology was an international finalist, and showed how mobile applications can allow citizen scientists to help document the location of endangered species. Overall, 10 high school students participated in a dorm-based summer camp from funds provided by the project. Two workshops on the topic were organized at international workshops.
