Geocent was recently invited to participate in the PlugFest event at the TechNet Land Forces Southwest conference in Tucson, AZ. PlugFest is an event in which various commercial entities provide a set of tools and expert support to participants in order to rapidly develop a solution to a problem revealed at the start of the event over the course of two or three days. By creating an environment in which software developers, solution providers, and subject matter experts work collaboratively to solve either a real or otherwise representatively real problem in a limited amount of time the event becomes a veritable proving ground for leadership, technology, and talent.
In essence PlugFest represents a proposed shift in the standard acquisition model from one of request → response → award. In such an environment awards are assigned based on limited information such as providers stating their technologies are standards-based and interoperable or technology firms proposing the development of a solution that should theoretically meet requested requirements. Instead PlugFest creates an environment where participants are required to prove themselves and their technologies in a rapid prototyping environment in order to provide a proof of concept to a problem at hand.
Development of the set of problems were closely modeled after the military’s Find, Fix, Finish engagement strategy. For the PlugFest the problems were defined as follows with no specification restricting solution methodologies or technologies
- Find – Locate a high value target based on intelligence information
- Fix – Determine the best method to move forces into a specified location
- Finish – Determine how best to strike a target based on in-situ factors.
The problems were revealed at the opening of the conference on Monday the 25th of March with presentation and judging of the participant’s solutions on Wednesday the 29th. Through a random selection the Geocent representative was assigned the first problem: Locate a high value target in the Jalalabad region utilizing any and all information and technologies available.
The problem at hand was extremely open-ended which was both daunting in its magnitude but freeing in its lack of restriction on delivery requirements. In this matter the benefits of the PlugFest environment immediately became evident as discussion with the present military subject matter experts narrowed the problem to a matter of increasing the probably of a target existing in a location at a known time. To this end the problem could be broken down into two easily manageable factors: information management and geospatial visualization.
By providing a method for information analysts to better research and manage intelligence information, relevant information could be used to build a more relevant information store or a potential target. Additionally, by providing a geospatial and temporal representation of the intelligence information an analyst should be able to better identify spatial and temporal patterns in order to more accurately develop a fix on the target.
As parts of the solution would likely entail logically different tasks within a larger intelligence workflow the tasks would be developed within separate workspaces while utilizing open standards to allow interoperability. To provide a customizable view of the solution the tasks would be developed as widgets deployed to the OZONE Widget Framework (OWF).
The design of the information management task was broken down into two logical sub-components: intelligence search and intelligence management. In essence an analyst would be able to research a specified target using the search component and collate the relevant data utilizing the management component. Therefore the skeleton for two OWF widgets were developed and deployed: an Intelligence Search widget and a Dossier Manager widget.
Intelligence Search Widget
Lacking any example intelligence data for the PlugFest event a focus was placed on “open source” data repositories in order to simulate a military analyst tasked with researching a specified target from various intelligence data sources. The implemented search mechanism would make Ajax requests to search endpoints implemented against a specified client-side interface with a translation to a standardized data format for unified management and result display within the widget. This would allow the integration of any number of intelligence data sources into a single search interface by simply implementing the defined interface and performing the data translation logic.
To simulate a historical set of intelligence data an interface was created to perform a free text query against the public search interface of Wikipedia. To simulate near real-time intelligence data an interface was created to perform a free text query against the public search interface of Twitter.
To allow for interoperability with other modules or widgets within the OZONE Widget Framework a communication channel was created using the OWF Eventing API. A supporting action was also created which allowed one or more elements of intelligence information in the defined standard format to be serialized into JSON and submitted to any elements subscribed to the created communication channel.
Figure 1: Intelligence Search Widget Data Workflow
While the data services used have the option of providing geospatial metadata pertaining to the result information it cannot be assumed that all items will be properly marked up with this information thus making geospatial visualization impossible. In the case where the search results do not contain this metadata a workflow was developed in order to discover relevant geospatial information using natural language processing combined with geotagging.
An XML and JSON interface local to the widget’s servlet was created utilizing Spring’s REST capabilities which accepted a string of text as input. The data was processed using Apache OpenNLP with the provided english models for Sentence Detector (en-sent.bin), Tokenizer (en-token.bin), and Location Name Finder (en-ner-location.bin) to extract any geographical or landmark based information. For each location that was located a request was then made toGeoNames (a public geotagging service) to determine the coordinates of each location. By using this workflow all search results in the widget may be properly assigned geospatial coordinates where relevant metadata is not provided.
Figure 2: Geotagging Workflow
Dossier Manager Widget
Individual pieces of intelligence information may not be of any use without being put into context with supporting information. In order to provide a mechanism by which an analyst would be able to collate and organize intelligence information on specified target the Dossier Manager widget was developed. This widget provided an interface to enable the easy creation and management of target dossiers whose contents could be populated from the Intelligence Search Widget.
The widget obtained data for the selected dossier by subscribing to the Intelligence Search Widget’s communication channel and appending the selected intelligence items to the target’s list of information. To increase usability the OWF Drag and Drop API was utilized to allow easy management of the data by simply selecting one or more items from the Intelligence Search Widget and dragging them to the Dossier Manager Widget.
This information was persisted on a per-user basis utilizing the OWF Preferences API allowing each analyst to create a customized overview or overviews of targets based on their personal preferences or situational requirements.
To allow for interoperability with other modules or widgets within the OZONE Widget Framework a communication channel was created using the OWF Eventing API. A supporting action was also created which allowed the dossier of intelligence information to be serialized into JSON and submitted to any elements subscribed to the created communication channel.
Figure 3: Dossier Manager Widget Data Workflow
The value in maintaining information dominance in the field of intelligence has always proven to be of the utmost importance when developing strategies or responding to emerging events. Additional benefits are gained when this intelligence can be visualized in a geospatial manner especially when fused with existing geospatial in situ data systems. From conversations with military subject matters experts it was derived that providing the ability to visually mash up this information for an experienced analyst would allow the identification of geospatial patterns, temporal patterns, or emerging points of opportunity. By providing this capability the end goal of increasing the probably of a target’s location existing in a known space and time was achieved.
Intelligence Visualizer Widget
With a known goal of providing a geospatial view of intelligence data fused with additional geospatial information an Intelligence the Visualizer Widget was created utilizing the OpenLayers dynamic mapping software. By utilizing OpenLayers any additional source which provided its information in an Open Geospatial Consortium (OGC) standards-compliant format (WMS, WFS, GML, etc.) could be viewed alongside intelligence data to provide additional analysis capabilities for the end user. Additional management capabilities were developed for the analyst to add these additional OGC data sources and persisted on a per-user basis for personal and situational customization.
The widget obtained data for the selected dossier by subscribing to the Dossier Manager Widget’s communication channel and creating geospatial markers with supporting information for each intelligence data element.
Figure 4: Intelligence Visualizer Widget Data Workflow
The PlugFest event offered a chance to work with technical experts from participating companies in order to quickly field their capabilities in a rapid development environment. Data provided by Esri in various OGC formats in order to simulate both real time and historical tactical information. Also, impressive teaming by IBM and ETCorp provided access to a real-time sensor system through a standards-compliant KML interface. Additionally,
By utilizing these capabilities the Geocent solution was further enhanced by allowing a military analyst to mash up intelligence data with tactical information and verify through any nearby elements of the real-time sensor systems thus providing increased capabilities to find and fix the theoretical high value target.
Upon reviewing the solution by PlugFest’s participating subject matter experts additional benefits were identified by identifying additional technologies that could provide data sharing capabilities and cross domain search capabilities with minimal integration costs.
Through a rapid development environment whereby subject matter experts exist alongside industry representatives and development resources a tightly iterative prototyping process could be achieved via a continuous cycle of requirements refinement, technology integration, prototype development, prototype demonstration, and product evolution through feedback.
By the conclusion of the two day event all participants had developed a demonstrable product that provided a solution to their assigned problems. As a result representatives were immediately able to validate claims of technology functionality, interoperability, and performance along with assessing participating company’s capability of providing software development services within requirement deadlines without the need to fully execute the traditional acquisition workflow.