Toward sensor-based context aware systems.

Area of Science:

• Computer Science
• Artificial Intelligence
• Sensor Networks

Background:

• Interpreting complex sensor data often requires contextual understanding beyond raw readings.
• Existing methods may struggle to reconcile diverse uncertainty types from multiple sensors.
• Business rules offer a structured way to represent operational contexts.

Purpose of the Study:

• To propose a methodology for sensor data interpretation that combines sensor outputs with annotated business rules.
• To develop a system that generates context-specific events with uncertainty levels.
• To demonstrate a feasible approach for context-aware decision-making using sensor data.

Main Methods:

• Sensor readings are interpreted to generate events with uncertainty labels.
• Contexts are represented as sets of annotated business rules.
• Uncertainty is reconciled by transferring it from events to business rules via Boolean predicates.
• Context rules and events are evaluated for decision-making.

Main Results:

• A context-reasoning engine was successfully connected to simulated heartbeat sensors.
• The system demonstrated the ability to identify the proper operational context using sensor outputs.
• Decision-making was effectively triggered based on the identified context and evaluated rules.

Conclusions:

• The proposed methodology provides a robust framework for sensor data interpretation and context-aware decision-making.
• Integrating sensor data with business rules offers a practical solution for complex operational environments.
• This approach enhances system adaptability and responsiveness by leveraging contextual information.