1. Marko Jurmu, MediaTeam Oulu 1 (22) Context-Aware Computing Overview and Case Studies Marko Jurmu MediaTeam Oulu Group University of Oulu Finland March 13 th, 2007

2. Marko Jurmu, MediaTeam Oulu 2 (22) Outline: Fact Sheet: Finland Motivation Definitions of Context Effects of Context-Awareness Example Application Domains CASE 1: Connectivity Management CASE 2: RFID Symbolic Tags CASE 3: Wearable Sensors Raised Issues References

3. Marko Jurmu, MediaTeam Oulu 3 (22) Fact Sheet: Suomi (Finland) Population: 5.3 mi Area: 330.000 km 2 (127.000 sq mi) Independent since: Dec. 6, 1917 Capital: Helsinki (ca. 1 mi) Home of: –Santa Claus –NOKIA –Linus Torvalds –Jari Litmanen –Teemu Selänne –Jari Kurri

4. Marko Jurmu, MediaTeam Oulu 4 (22) Motivation Traditional computer systems execute blindly –Pre-defined application logic –Static user interfaces that require explicit attention New kind of computing –Mobility is becoming a norm –Applications execute in less static environments What can be done? –Enable computer systems to sense their surroundings –Encode changes in the surroundings to reflect to the application behavior

5. Marko Jurmu, MediaTeam Oulu 5 (22) Motivation cont’d But: –Coding explicit reflections into applications is laborous –Applications do not necessarily understand all changes –Applications may not be able to collect all the necessary information We need: –Supporting platforms that facilitate application execution –Ways to gather disparate information and provide it in a unified way –Mechanisms to provide information without loss in semantics These are the main research problems of context-aware computing

6. Marko Jurmu, MediaTeam Oulu 6 (22) Definitions of Context Schmidt et al. [1]: –Context is a series of transient hierarchical feature spaces –Very systematic definition

7. Marko Jurmu, MediaTeam Oulu 7 (22) Definitions cont’d Dey [2]: –”Context is any information that can be used to characterize the situation of an entity. An entity is a person, place, or object that is considered relevant to the interaction between a user and an application, including the user and the application themselves.” –Very loose definition, reflects well the common paradox in context- aware computing: Applications can never be made aware of all possible context In practice, the subset of context utilized by applications is strongly dictated by the sensing capabilities of underlying computer system

8. Marko Jurmu, MediaTeam Oulu 8 (22) Definitions cont’d Dey’s definition of context-awareness [2]: –”A system is context-aware, if it uses context to provide relevant information and/or services to the user, where relevancy depends on the user’s task” –Important contrast: Previous discussion was strongly related to providing information that enables dynamic changes in the execution of the application According to this definition, application is also context-aware if it can use context to filter the information presented to the user

9. Marko Jurmu, MediaTeam Oulu 9 (22) Levels of Context In their work [1], Schmidt et al. define 4 levels of context: –Raw sensor data Readings acquired from individual sensors –Cues Abstracted sensor data, possibly temporally aggregated –Context Description of the current situation on an abstract level –Scripting Allowing applications to benefit from context through simple scripts

10. Marko Jurmu, MediaTeam Oulu 10 (22) Effects to Application Logic Context-aware applications need to be designed differently from static applications: –Agile behavior on the face of rapid changes –Platform independence is required –Logic is distributed to better reflect the variance in the computing environments The reflection to context changes can be encoded in several ways: –Definition of explicit rules and conditions –Increasing / decreasing application functionality –Dynamically maintaining the FSMs that control the application

11. Marko Jurmu, MediaTeam Oulu 11 (22) Effects to Interaction Traditional interaction: –Based on explicit user input –Static UIs with a certain conceptual models –User overwhelmed with unnecessary information Situations worsens with ever-increasing information Context-aware interaction: –Based on combination of implicit and explicit I/O Implicit interaction does not require additional cognitive load –UIs that filter information and lessen the explicit controlling –More intuitive interaction models for users can be enabled

12. Marko Jurmu, MediaTeam Oulu 12 (22) Example Application Domains Location-aware applications: –Map-based guidance –Focused advertising –Weather / traffic services Network-aware applications: –Video streaming –VoIP –File management operations Combinations of these and other contexts: –Basically the list is endless, the following presents 3 enablers

13. Marko Jurmu, MediaTeam Oulu 13 (22) CASE 1: Connectivity Management Background: –Currently, wireless networking appears through heterogeneous overlapping access networks –Mobility of the users is causing constant changes to the configuration of active radio interfaces –Connectivity can be based on infrastructural or ad-hoc links –Networking applications in the mobile device need coherent mechanisms to cope with networking conditions

14. Marko Jurmu, MediaTeam Oulu 14 (22) CASE 1 cont’d Solution [3]: –Introduce a persistent networking interface for applications –Applications only see channels, and can specify different policies for network utilization

15. Marko Jurmu, MediaTeam Oulu 15 (22) CASE 2: RFID Symbolic Tags Background: –Invoking specific actions from mobile phone requires the navigation of hierarchical menus –Can be very time-consuming, and does not guarantee that all action-related parameters are entered correctly –All in all: Highly explicit UI

16. Marko Jurmu, MediaTeam Oulu 16 (22) CASE 2 cont’d Solution [4]: –Add RFID tags with intuitive symbols to the environment –Tag has action-related parameters encoded, so menu-based setting is not required –Still, users only see symbolic actions that can be invoked by touch

17. Marko Jurmu, MediaTeam Oulu 17 (22) CASE 3: Wearable Sensors Background: –Computer systems need information regarding the activities of the user to be able to support them –Having basic context like location is not necessarily enough –Activity is also correlated with artifacts being currently utilized

18. Marko Jurmu, MediaTeam Oulu 18 (22) CASE 3 cont’d Solution [5]: –Equip users with wearable sensors –Additional sensors in artifacts give information on usage

19. Marko Jurmu, MediaTeam Oulu 19 (22) Raised Issues User control –Context-awareness takes control away from users –What is the right amount of user control? User must feel that he/she is in control of the system –Two extremes: Application asks everything from user (not very feasible) (certainly not very feasible) Privacy –Context-awareness requires the acquisition and storaging of information related to users and their surroundings –How to ensure correct management and avoid leakages? –How to avoid the collection of sensitive information?

20. Marko Jurmu, MediaTeam Oulu 20 (22) Future Trends From application point-of-view, mobility is turning multidimensional –Over administrative domain boundaries –Over device boundaries –Over social boundaries Living environments are increasingly saturated with intelligent electronics –Available services increase –Digital and physical worlds increasingly overlap Challenges lie ahead especially in dynamic session management between environments and mobile users

21. Marko Jurmu, MediaTeam Oulu 21 (22) References [1] Schmidt, A., Beigl, M. & Gellersen, H.-W. (1999), There is more to context than location, Elsevier Computer and Graphics Journal 23(6): 893-901. [2]Dey, A. (2000), Towards better understanding of context and context-awareness, CHI 2000 Workshop on the What, Who, Where, When, Why and How of Context-Awareness. [3]Sun, J., Riekki, J., Jurmu, M. & Sauvola, J. (2005), Adaptive connectivity management middleware for heterogeneous wireless networks, IEEE Wireless Communications 12(6): 18-25. [4]Riekki, J., Salminen, T. & Alakärppä, I. (2006), Requesting pervasive services by touching RFID tags, IEEE Pervasive Computing 5(1): 40-46. [5]Pirttikangas, S., Fujinami, K. & Nakajima, T. (2006), Feature selection and Activity Recognition from Wearable Sensors, Proc. International Symposium on Ubiquitous Computing Systems (UCS2006), Seoul, Korea, Oct. 11-13, 2006, pp. 516-527.

22. Marko Jurmu, MediaTeam Oulu 22 (22) Contact: marko.jurmu@ee.oulu.fi www.mediateam.oulu.fi www.oulu.fi http://en.wikipedia.org/wiki/Finland