Group-Context-Aware Mobile Applications

Handheld Devices , Tactical Systems Add comments

By Marc Novakouski,
Member of the Technical Staff
Research, Technology & System Solutions

Marc NovakouskiOur modern data infrastructure has become very effective at getting the information you need, when you need it. This infrastructure has become so effective that we rely on having instant access to information in many aspects of our lives.  Unfortunately, there are still situations in which the data infrastructure cannot meet our needs due to various limitations at the tactical edge, which is a term used to describe hostile environments with limited resources, from war zones in Afghanistan to disaster relief in countries like Haiti and Japan. This blog post describes our ongoing research in the Advanced Mobile Systems initiative at the SEI on edge-enabled tactical systems to address problems at the tactical edge.

At the tactical edge, the people that need the information the most—warfighters, first responders, or other emergency personnel—depend on timely and valuable information to perform their tasks, or even survive.  Unfortunately, the access to the information they need can be extremely hard to achieve, for the following reasons:

  • information overload stemming from too much information, coupled with an inability to locate truly vital information
  • information obscurity due to a lack of awareness of the available information, aka “you don’t know what you don’t know”
  • resource scarcity manifested as insufficient bandwidth, central processing unit (CPU) power, battery power, or even attention span to get the needed information and continue to process, exploit, and disseminate it for as long as needed

The remainder of this posting describes how we are tackling the information overload and information obscurity aspects of this problem by developing context-aware mobile applications.

A Different Approach to Context-Aware Mobile Applications

Context awareness in the mobile environment is not a new field of research.  Most mobile devices come preloaded with applications that use location or time to account for user context. There is certainly no shortage of similar applications available for download.  We decided, therefore, to explore alternative sources of data that would not only push the limit of what could be done with user context, but also focus on the extremely challenging environment at the tactical edge.

Our “eureka” moment came when we realized that when warfighters or first responders are at the tactical edge, they are almost never operating alone.  As a result, the most important contextual information to warfighters or first responders is the context of the people in the group, and how they relate to that context.  This realization drove us to explore group context-aware mobile applications.  These applications would, if built correctly, first consider individual user context and then relate that information to the group context, thereby helping users understand both their own state, as well as the state of the group in which they participate.

Group context-aware mobile applications clearly have value at the tactical edge.  For example, warfighters are well served by having access to positions of friends and foes on the battlefield (position data being a simple case). They could also benefit from supportive applications that monitor resources, such as food, ammunition, or vital signs.  With sufficient data and processing power, these applications could even use historical trends to determine dynamically if a squad is walking into a possible ambush situation.

In less deadly (yet still hostile) environments, such as tsunami disaster areas, the ability to share information about resource needs, dangerous situations, or health emergencies in a structured way can also be valuable.  Such applications could tailor information to managers, construction workers, doctors, and other emergency personnel to help coordinate an effective emergency response.

An extensive literature review on context awareness yielded relatively little research on the topic of group context.  Much of the prior work cites the basic context model developed by Anind Dey, but does not expand the model past the individual, choosing instead to tailor the model to a particular domain.  Our research project, called Information Security to the Edge (ISE), explores the structure, applications, and implementation of a context model that includes group information.  We have constructed a prototype application on the Android platform that implements the essential components needed by group context-aware mobile applications, as discussed next.

App Architecture – Logic and Data

The ISE prototype application follows the common model-view-controller (MVC) pattern, which decomposes an application into the following parts:

  • The model is the data. This data is the information processed by the application.  For example, the words typed by the user into a word processing application are data.
  • The view is the user interface. In the case of a word processing application, the view would be the buttons, menus, scroll bars, and other visual effects provided by the application to help a user write a document. 
  • The controller is the logic. In the case of a word processing application, the controller would be the rules the application uses to save, present, filter, and otherwise modify the text.  The function provided by each button or menu item can also be part of the controller.

Consistent with the MVC pattern, the ISE prototype has a central control mechanism (which forms the “brains” of the application) that manages data flow through the application.  In practice, this means that the central controller coordinates data flow and processing through the following primary application elements:

  • The context engine is the central processor for all context information used by the application.  As device sensors report new data and applications on external devices send data to the local application, all data is passed through the engine so that new events are detected as they occur.  For example, if an external user sends their GPS coordinates that indicate they are within 100 feet of a warfighter, then the device can alert the warfighter to their presence.  Expanding on this concept, if a group task must be performed but everyone is working individually on their own tasks the local device can monitor task status and user position and report to the leader when all group members are ready and close by so the group task can be performed.
  • The sensor manager accepts data from sensors that reside upon the mobile device.  A typical smart phone contains position sensors, movement sensors, and in some cases, light and proximity sensors.  The application captures data from these sensors and passes it through the sensor manager.  The sensor manager enables the sensors and controls their sample rate, so that the application can tailor usage to the situation and avoid overwhelming the system.
  • The communications manager acts as the gateway to all external communications within the system.  This gateway currently includes Bluetooth and TCP/IP communications, but can be expanded to include other communication mechanisms that are available to the device.  Any messages to and from users on other devices are passed through the communications manager.

The sensor and communications manager architecture consolidates all sensor and communication concerns into a single location.  This consolidation approach enabled us to build a standardized interface that simplifies integration an arbitrary sensor (for example, a radiation sensor) or an arbitrary communication mechanism (for example, a line-of-sight radio that communicates with UAVs) with the application.  We tested this feature through a collaboration with Joao Sousa of George Mason University. This testing resulted in the development of an alternative communication mechanism that integrates with the prototype with only a few weeks of effort, instead of months or years.  We anticipate leveraging these standardized interfaces to collaborate with a variety of external groups and organizations as new sensor technologies and communication mechanisms become available.

  App Architecture – User Interface (UI)

The ISE app, through the use of Android UI screens called Activities, reflects the view part of the MVC Pattern. There are currently only three supported UIs in ISE:

  • User: Allows users to look at the people with whom they are or can be connected, as well as the context data associated with each person.
  • Task View: Allows users to create their own tasks, receive updates about other users’ tasks, and mark their tasks complete or incomplete. We are expanding the task view to include tasks, with main tasks and subtasks under them. Ultimately, we will develop a capability that displays complex missions in an intuitive manner.
  • Alerts View: As events occur, some will automatically appear in the alerts view along with a list of the considerations the context engine has identified as items of importance for users. The alerts presented will be tailored to the needs and context of individual users.

We are upgrading the ISE architecture to support any UI that subscribes to standardized updates from the data services.

Challenges

One challenge we face involves accounting for the lack of network infrastructure.  In particular, limited bandwidth exists for the available communication channels.  We are building atop of communication capacities that other organizations are field testing in Afghanistan to tailor our solution to practical field situations. 

A second challenge involves providing warfighter access to backend data sources.  Soldiers told us that important information is available in such sources, but they can’t readily find the relevant information. Moreover, they can’t access the database in the field.  Other Advanced Mobile Systems work is investing ways to provide access to critical data through the use of cloudlets.

A third challenge involves reducing the user’s cognitive load by limiting the amount of interaction and attention required of the user. Residents in a metropolitan area can use their smart phones without undue concern for being in a distracted state, as long as they are not engaging in tasks that demand undivided attention. A soldier in Haiti, on the other hand, must be cognizant of crumbling buildings, while a warfighter on the ground in Afghanistan might need to digest information while taking enemy fire. Our goal is to use hardware that allows the warfighter to capture and process information seamlessly, without sacrificing valuable time and resources.

We are also addressing the challenge of resource scarcity.  Resources are limited at the tactical edge and warfighters are typically limited to the power and bandwidth of whatever devices they can carry.  We are therefore exploring resource optimization based upon our expanded model of context. For example, if a warfighter’s assignment involves driving through a known safe area, it may not be necessary for the smartphone to activate the GPS capability.  By optimizing the system to use sensors only when needed, warfighters can save battery power, CPU cycles, and communication bandwidth that can be used to support other mission-critical needs.

Finally, our work will not have the desired impact if we cannot meet the challenge of relevance. Warfighters made it clear to us that if a device or application is not directly useful to their immediate task, it will be ignored.  In any given day, a warfighter in Afghanistan may be asked to determine if a particular individual is a threat, sweep a village to establish identities of residents, deliver food to children, or check for a weapons cache. These different missions impact the type of information that interests soldiers and the type of information a software application should consider.  Solving this problem requires a deep understanding of the needs of soldiers and the missions in which they engage.  We are leveraging this domain knowledge so our ISE application can tailor information processing to a particular mission, thereby ensuring relevance to the current mission and the ability to change mission parameters as needed.

Looking Ahead

The ISE prototype is just one part of our strategy to address the problems of information overload, information obscurity, and resource scarcity.  The Advanced Mobile Systems initiative is also engaged in the Edge-Enabled Programming project as well as the Resource Optimization for Mobile Platforms at the Edge project.  Each project attacks the three problems of information overload, information obscurity, and resource scarcity from different perspectives. We intend to integrate each project together after they have matured, thereby providing an end-to-end solution to warfighters and first responders at the tactical edge.

Additional Resources

For more information on the MVC pattern, consult the Documenting Software Architectures: Views and Beyond and the Pattern-Oriented Software Architecture: Volume 1 books.

Share this

Share on Facebook Send to your Twitter page  Save to del.ico.us  Save to LinkedIn  Digg this  Stumble this page.  Save this page on your Google Home Page 

2 responses to “Group-Context-Aware Mobile Applications ”

  1. Parker Green Says:
    I have been on the search for some type of mobile capture application. This isn't that type of app is it?
  2. Marc Novakouski Says:
    It depends on what you mean by mobile capture. This application will capture and record data from on-board sensors, as well as user input and shared contextual information from group members. If you are referring to audio or video capture, then I can say that while the app does not currently support that capability, we are planning on integrating those sources as additional on-board sensor data sources. Can you describe a scenario for the capabilities you are looking for?

Add Comment


Leave this field empty: