Daily Note : Situated Computing and Intelligent Systems

These notes are a summary of concepts presented in “Towards Situated Computing.”

R. Hull, P. Neaves and J. Bedford-Roberts, “Towards situated computing,” Digest of Papers. First International Symposium on Wearable Computers, Cambridge, MA, USA, 1997, pp. 146-153, doi: 10.1109/ISWC.1997.629931.

  1. Situated Computing Systems
    • Computing devices detect, interpret, and respond to the user’s local environment
    • Data elements
      • Personal health
      • User identity
      • Companions
      • Location
      • Time
      • Computing resources
      • Physical environment
    • Applications and Features
      • Augmented reality
      • Localized information
      • Context-based retrieval
      • Remembrance agents
      • Interface interactions
      • Monitoring capabilities
      • Opportunistic task support
  2. Technology Components
    • Radio communication marker tags
    • Sensors feeding data to service APIs
      • Data fusion and interpretation
      • Event delivery to applications
      • Query interface for situational data
  3. System Design
    • Combine and interpret sensor data to provide situational information
    • Post events to applications when situations change
    • Provide a query interface for current situational data interrogation
    • Monitors sensors and sends event-based notifications to applications
    • Operates similarly to an OS service monitoring user actions
  4. System Elements
    • Transparency
      • Transparent access to situational information when sensors are available
    • Dimensional Independence
      • Independent management of situational dimensions
    • Low Latency
      • Prioritize processing situational events for timely responses
    • High Throughput
      • Handle many concurrent events efficiently
    • Extensibility
      • Support dynamic system configuration changes
  5. Interpretation Challenges
    • Provide a stable and abstract view of the user’s situation
    • Handle noisy, incomplete, or incorrect sensor data
    • Address redundancy with multiple sensors signaling similar aspects
    • Leverage data fusion for diverse sensing evidence
    • Avoid brittle heuristics that may produce erroneous inferences
  6. Interactions with Situated Computing Systems
    • Use situational-remembrance agents for cross-referencing multimedia captures in related contexts
    • Users often need to specify or configure their particular situation for applications