Privacy-based Analytics for Weather and Radiological Emergencies

The Department of Homeland Security (DHS) is looking for solutions that not only provide enhanced security and safety by leveraging sensors found on the Internet of Things (IoT) but also suppress data that facilitates the accumulation, processing, and characterization of information that make it attributable to individuals. To address these needs, RAM Laboratories proposes to develop Emergency Response Watch (ER-Watch), a suite of solutions that discern important events and actionable information for first responders with privacy services that suppress user-specific information. This tool consists of apps that extract specific readings from sensors prevalent on commonly used mobile devises, such as phones and tablets. The apps then generate event notifications that are tagged with Pedigree and Provenance information such as location, timestamp, and identifier information. ER-Watch uses back-end cloud processing to determine "watch" and "warning" regions for certain events by using analytics and data clustering algorithms. The Phase I project focuses on tornado warnings and warnings of radiological events. A key facet of the ER-Watch apps and processing are their incorporation of Privacy by Design (PbD) processes within its design. These PbD principles are used to (1) obfuscate device/user identifiers through the use of pseudo-randomization, (2) separate authentication and event analytical processing mechanisms to separate identify information from raw and derived sensor, and (3) use event processing to hide precise location and time tag and thwart track reconstruction activities. The Phase I project will be rapidly prototypes using existing apps previously developed by RAM Laboratories for sensor extraction and messaging.