Edgar Marko Trono

DTN MapEx: Disaster area mapping through distributed computing over a Delay Tolerant Network

Disaster area map generation and sharing are critical to disaster response operations. In post-disaster contexts however, cloud-based mapping services and data may be unavailable because of network challenges. Disruption Tolerant Network (DTN) architectures have been proposed for data sharing in challenged networks. However, map generation may be too complex for individual DTN nodes given their limited computing resources. To generate and share maps of disaster areas, we present DTN MapEx, a distributed computing system for mapping that operates over a DTN. DTN MapEx distributes disaster map data and map generation tasks to multiple nodes to minimize individual computational loads. In the system, responders and volunteers act as mobile sensing nodes. They log the GPS traces of their traversed paths and collect disaster area map data such as the coordinates, images, and assessments of points-of-interest. The mobile nodes then route their collected data and a task request through the DTN to pre-deployed, fixed Computing Nodes. The Computing Nodes aggregate the data to generate a map and opportunistically route it back to the network. To reduce complexity, mapping tasks and data are divided amongst Computing Nodes based on their current computational load. Computing Nodes periodically update the DTN about their current loads. Mobile nodes use these updates in deciding where to allocate their task requests and data. In this paper, we present the design of DTN MapEx and perform initial evaluations on its feasibility in disaster scenarios.

Disaster area mapping using spatially-distributed computing nodes across a DTN

Disaster area mapping is critical to guiding evacuees to safety and aiding responders in decision-making. During disasters however, Cloud-based mapping services cannot be relied upon, because network infrastructures may have been damaged. In this study, we propose a disaster area mapping system that functions under challenged-network environments in a disaster area. The system infers a pedestrian map with walking speed information from data gathered by civilians and responders with mobile devices. To generate the map, the system addresses the following challenges: how to collect disaster area data, how to share data without continuous end-to-end networks, and how to generate maps without Cloud-based mapping services. First, the system leverages human mobility to collect disaster area data. Civilians and responders with mobile devices function as sensor nodes and log their GPS and velocity traces while moving based on the Post-Disaster Mobility Model. Second, the system uses mobile devices to establish a Delay-Tolerant Network, through which nodes opportunistically share data. Finally to generate the map, the collected data are routed to Computing Nodes: devices with more computational resources than mobile devices that are spatially-distributed across the disaster area. The Computing Nodes infer the map from the data and share it with evacuees. Through experimental evaluations and computer simulations, we found that the system significantly decreases the time required to generate and deliver a map to an evacuee, compared to a case without the system. Furthermore, the overall reduction in time increases as the size of the data required to generate the map and the number of DTN nodes increase.

Milk Carton: A Face Recognition-Based FTR System Using Opportunistic Clustered Computing

Family Tracing and Reunification (FTR) is the process whereby families separated by disasters are reunited. Current FTR systems use either inefficient paper-based forms and notice boards or digital registries that need the Internet, which may be unavailable during disasters. In this demonstration we present Milk Carton: a system that aids in FTR. Milk Carton creates a registry containing evacuee records. To find separated persons, Milk Carton uses Eigenfaces face recognition to match queries with existing records. Milk Carton uses a clustered architecture of Computing Nodes to handle data storage and execute the Eigenfaces algorithm. To operate under challenged-network environments, Milk Carton uses response patrol vehicles as data ferries to deliver data. In this demonstration, we show how Milk Carton uses Eigenfaces to locate separated persons and how data ferries and Computing Nodes function.

Generating pedestrian maps of disaster areas through ad-hoc deployment of computing resources across a DTN

Generating pedestrian maps of disaster areas is an important part of response operations. Maps aid responders in decision-making and show routes that lead evacuees to refuges. However, disasters can damage communication infrastructures, rendering Cloud-based mapping services inaccessible. Responders resort to paper maps, which are difficult to share and cannot recommend routes. In this study, we present a digital pedestrian map generation system for disasters. To realize the system, we addressed these challenges: (1) how to collect the required data and generate the map without Cloud-based computing resources, (2) how to share messages within the system without continuous, end-to-end networks, and (3) how to balance the load of map inference tasks. For (1), GPS traces are collected by responders exploring the area. Then, collected data are sent to Computing Nodes: commodity workstations that are deployed in the disaster area, for processing. For (2), the system establishes a Delay-Tolerant Network that uses Epidemic Routing to communicate across shorter-ranges and uses response vehicles as data ferries to communicate across longer-ranges. For (3), we propose a load balancing heuristic, which uses ferry route timetables and statistical information about the load of Computing Nodes to determine how to offload map inference tasks. We evaluate our system through experiments and simulations and show that it decreases the time needed to generate and deliver pieces of the map by approximately two hours in an extreme case with large quantities of data have to be processed.

Milk Carton: Family Tracing and Reunification system using Face Recognition over a DTN with Deployed Computing Nodes

During the recovery period after disasters, Family Tracing and Reunification (FTR) is the process by which separated family members are reunited. Traditional FTR methods rely on paper-based registries and notice boards, which cannot automatically match missing person queries with existing records and cannot be efficiently disseminated. Furthermore, lost children or people with disabilities may not be capable of supplying the text-based personal information required by registry forms. Finally, current digital FTR systems require the Internet for data delivery and storage, which may be unavailable during a disaster scenario. To overcome these limitations, we propose the Milk Carton FTR system. Milk Carton uses the Eigenfaces face recognition algorithm to automatically match missing person queries with existing records, without requiring text-based personal information. The system uses Computing Nodes, commodity devices deployed in evacuation centers, to handle record and query creation, data storage, and matching. To handle data delivery without the Internet, Milk Carton leverages response team vehicles as data ferries. The ferries store, carry, and forward records and queries across the system. In this study, we present the design of the Milk Carton system and initial performance evaluations.

RecureShare — Internet-less application distribution mechanism for internet-less emergency communication systems

Currently, many studies are being done on smartphone-based disaster communication systems. However, most existing systems assume that all required applications are installed in the smartphones of all participants before the disaster occurs, which is unrealistic in actual disaster scenarios. If an Internet connection is unavailable, obtaining the applications necessary for communication becomes difficult because most applications are provided through the Internet by platforms such as Google Play. In this paper, we propose a mechanism for distributing applications and present its implementation. Our method uses the default Web browser and tethering function of the Android operating system. These, combined with a Web server embedded in our application, enable Internet-less application distribution. To evaluate the feasibility of our proposed system, we conducted an experiment with four participants who are unfamiliar with our application. Results show that the participants understood the system and were able to obtain the application without an Internet connection within a short span of time.