Is Context-Aware Computing Ready for the Limelight?
by George Lawton
Context-aware computing — which leverages information about users and their environments to improve the interactions among them — has been around for almost two decades. However, it has been implemented commercially only in limited applications, such as those using location information to find nearby friends or stores.
This has occurred mainly because of a lack of relevant standards, few devices with the capabilities necessary to perform context-aware computing, and limited sources of context-related information to draw on.
Now though, context-aware computing is ready to take off, due to significant improvements in social networking, mobile technology, smart phones, and sensor networks, said University of Colorado associate professor Rick Han.
In addition, researchers have become better at combining location-based, presence, social-network-profile, and other data to determine a user's context, said William Clark, research director at market research firm Gartner Inc.
Because of this, vendors and researchers are beginning to release more advanced context-aware computing applications such as recommendation engines that run on mobile phones, and better customer-service and information-management tools.
However, the technology still faces technical and marketplace challenges, such as those related to privacy and data integration.
Some of the early ideas behind context-aware computing came from ubiquitous-computing research conducted at Xerox's Palo Alto Research Center (PARC) in the early 1990s. This cutting-edge research required customized and expensive hardware, software, and networks.
Today, though, many of these formerly-advanced technologies are affordable and widely available. "The price of hardware has come down considerably. And everyone with a smart phone has a context-aware device," said Carnegie Mellon University associate professor Anind K. Dey.
In the early 1990s, putting sensors on devices wasn't easy, according to Bo Begole, a principle scientist and manager of PARC's Ubiquitous Computing Group. Now, he said, that is simple.
Over time, various vendors have introduced CA tools and hardware, such as Cisco Systems' Collaboration in Motion, Avaya's Aura, Rearden Commerce's Rearden Personal Assistant, several AeroScout products, several Appear Networks products based on its Appear Context Engine, and RES Software's RES PowerFuse.
Goals and Research Focus
Context-aware computing's goal is to let systems react to users based on their environments.
For example, a user could employ a context-aware system to look for a movie and theater. The application would combine information about the location of the user and nearby theaters, the users' film-preference history, the current time, and upcoming movie schedule to find an appropriate nearby show.
Context-aware computing can also improve business processes such as sales, inventory, scheduling, and purchasing by tailoring the way an application presents information to customers and employees, formulating suggestions, and automating some parts of the decision-making process.
For example, a sales application might prominently list the products that customers purchased in the past to make it easier for them to find what they want, or it might suggest sales-call opportunities based on a customer's buying history.
The technology could also improve customer-service interactions by providing company representatives with information about customers and even suggesting new products they might want.
Context-aware-related research leverages work in various areas including sensors, social networking, semantic analysis, location awareness, activity recognition, and machine learning.
How It Works
Context-aware systems make decisions and provide information based on multiple sources of information about users and from applications — acting as sensors — that gather data from users and their surroundings.
A basic system includes a contextually responsive application and hardware elements such as PCs, sensors, and switches.
Some of the most useful sources of contextual data include location (captured, for example, from a GPS system), identity (from information a system gathers about a user), activity (such as from a smart-phone-based to-do list), and time (from the system clock).
A smart phone is well-suited to context-aware computing because it brings together multiple data streams, including those related to user location and movement, and communications history.
Today, context-aware systems can also gather information from social networks.
Once they have acquired data, systems must understand the context that the data represents and determine its relevance to user queries and activities. Semantic technologies make it easier for systems to sort through, organize, aggregate, and interpret data streams.
Ultimately, context-aware systems must identify what actions to take based on the information. They can use expert systems to make decisions by determining how an organization's rules apply to the semantic analysis of context-related data.
Organizations have employed context-enriched services on a limited basis.
For example, airlines use the technology to provide relevant information — such as weather conditions at destinations — to customers via the Web, telephone, or applications such as e-mail, said Venkatesh Krishnaswamy, senior manager with Avaya's Emerging Technologies and Products Division.
Retailers have deployed context-aware systems to track a customer’s presence on their website and to start an interaction — such as trying to sell a specific product — based on context like the visitor's activity or navigation patterns.
Corporations use the technology to work with employee information such as communications patterns and collaboration histories to, for example, present the people most frequently contacted at the top of an employee's personalized contact list.
Some corporations use context-aware computing to optimize and economize their companywide software-configuration process, added Jim Kirby, RES Software’s president for the Americas. RES tools analyze users' context to automatically recognize which software they work with, and then configure their workspaces with only the necessary applications. This reduces software-licensing overhead.
The New Context
The rise and increased availability of location-based technology and social networking have helped make context-aware computing easier to implement.
Systems can acquire considerable relevant data about users from their locations, as revealed by GPS readings or geolocation sensors.
"The development of standardized interfaces for exchange of location data between applications is a key enabler," said Avaya's Krishnaswamy.
And, he added, "Several start-ups are beginning to offer services such as semantically rich mapping data, real-time traffic information, location-aware advertising or promotions, and location-aware social networking."
According to the University of Colorado's Han, "The content on social networks can provide a whole new set of context that was not available before."
If users allow a context-aware application to access their social networking data, the system could then collect relevant information.
The technology could also help implement policies regarding presence, which describes a user's willingness and availability to communicate. Context awareness could help automatically set users' presence settings based on their preferences and activity, as well as the time of day.
As the technology grows and captures more data, privacy could become a concern. Efforts to improve authorization, authentication, and accounting could increase trust in context-aware systems, which could encourage adoption.
Advances in networks, mobile hardware capabilities, social computing, service-oriented architecture, and unified communication will make building and using context-enriched services easier.
Context-aware computing can help businesses by increasing efficiency, thereby reducing costs and improving workforce effectiveness.
By gathering relevant information about customers, the technology also improves the sales, product development, and service processes. This would let companies more effectively target, attract, and retain new customers; create and customize new products and services; and expand into new markets and geographical areas.
Information security and privacy applications could also use contextual data to function more effectively.
In July, Radiant Logic released Context Browser, a Web-based application that lets users search contextually across structured information stored in different places.
Globys recently launched the Mobile Occasions contextual marketing service for cellular-phone providers. It combines billing and demographic information, as well as other data, to target advertisements to customers.
Ryerson University's student-based Digital Media Zone and Appear Networks are testing an application that users with disabilities could download to an Android smart phone to help them navigate the Paris subway. The Mobile Transit Companion gathers contextual information about train schedules, users' destinations, and their special needs, and presents applicable travel instructions in a way that is appropriate for their disability. For example, a blind person would receive audio instructions.
Cisco is working on the Collaboration in Motion initiative. This project would integrate location, activity, behavior, network traffic patterns, and other types of information from the company’s routers, as well as other communications and networking products, to bring context-aware technology to mobile applications.
AeroScout has implemented Cisco's platform commercially in applications and handheld devices that gather contextual information about a patient and hospital resources, and enable healthcare workers to quickly find the type of equipment and supplies needed.
AeroScout uses the same technology in products for the manufacturing, logistics, and transportation industries.
A key element in this implementation is the ability to leverage information from multiple sources, explained Steffan Haithcox, AeroScout Senior Director of Marketing. For example, the system gathers sensor data from patients via Wi-Fi and tracks equipment via GPS or RFID technology.
Context-aware technology can add cost and complexity to applications.
One problem the approach faces is that many organizations acquire context-related data from different sources, many of which use different data formats. This creates potential data-access and -integration problems
Moreover, the ability to aggregate large amounts of data could make it easier to uncover private information about people, said Han.
Some advanced context-aware services require considerable time and computational power to process large amounts of data, said PARC's Begole.
According to Carnegie Mellon's Dey, proponents must better prove the technology's value. "While I think we’ve made considerable progress in building more compelling applications, I don't think we've done a good job explaining why adding context to an application is a valuable asset." And if context doesn’t noticeably improve users' experience with devices or applications, they're likely to simply disable the capability, he added.
In the long run, context — with its promise of intelligently improving the way users work with devices and applications — could be just as important as Internet search technology, said Gartner's Clark.
Context could become a source of differentiation for software and handset vendors, mobile network operators, and communications vendors.
However, said Begole, a privacy-related debate will occur over the type of information that context-aware applications should be able to access. There is a risk of backlash if consumers don't feel protected, he said. "People are not sure if the benefit of exposing that much personal information outweighs the risk," he explained.
Nonetheless, Hossein Rahnama, research director of Ryerson University’s Ubiquitous and Pervasive Computing Lab and Appear Research Labs' director of innovation, predicted good things for context-aware computing.
The technology will get easier as the tools and platform mature, he explained. As this occurs, he said, software vendors won’t have to develop all tools and applications components on their own.
For this and other reasons, he concluded, "The future is going to be context aware."
George Lawton is a freelance technology writer based in Monte Rio, California. Contact him at firstname.lastname@example.org.
For more on context-aware computing, see Computing Now's August 2010 theme.