TL;DR
This paper introduces LOCK, a novel Location-Based Access Control (LBAC) system that offers high accuracy and ease of use. By leveraging autonomous ultrasound positioning devices and a sophisticated access control engine, LOCK precisely defines secure zones and accurately judges access authority, achieving centimeter-level resolution without manual setup.
What Problem Does This Paper Address?
In our increasingly connected world, digital access is vulnerable to unauthorized intrusions from anywhere. Location-Based Access Control (LBAC) aims to mitigate this by using a user's physical location as a security factor, restricting access to specific "secure zones." However, existing LBAC systems often struggle with either coarse location judgment resolution or require significant manual effort for setup and calibration. This paper addresses the need for an LBAC system that is both highly accurate and user-friendly, capable of precisely defining and enforcing access policies based on location.
Methods
LOCK employs a two-phase approach:
- Offline Phase:
- Measurement-Free Calibration (MFC): This novel method automatically calibrates autonomous ultrasound positioning devices. It transforms their relative positioning data into an absolute coordinate system without requiring manual measurements of the room or device placement. This is achieved by using feature points (e.g., room corners) and solving a system of equations derived from coordinate mapping and distance equalities.
- Coherent Secure Zone Fitting (CSZF): To precisely define secure zones, CSZF uses a polygon fitting approach. A user moves a tag along the boundary of the desired secure zone, and the system collects positioning data. A slope-based subsequence division algorithm is then used to identify edges, followed by line fitting to determine the polygon's vertices. This method compensates for disparities between manually defined zones and those perceived by the positioning devices, improving accuracy.
- Online Phase:
- Round-Trip Judgment (RTJ) Algorithm: This algorithm efficiently determines the geographical relationship between a client's real-time position and the defined secure zones. It leverages the idea that if a point is inside a polygon, a trip along the polygon's boundary, originating and ending at the same vertex, should traverse the coordinate system's quadrants in a specific way. This allows for fast, O(n) complexity, online access decisions.
Datasets
Not discussed in the paper. The system was evaluated using a prototype in a lab environment.
Key Insights and Ideas
- Calibration without Manual Effort: The MFC method is a significant contribution, removing the tedious and error-prone manual calibration steps common in many indoor positioning systems. This directly addresses the "easy-to-use" requirement.
- Bridging the Gap in Secure Zone Definition: CSZF is crucial for aligning the system's perception of a secure zone with the user's intent. By using the same positioning noise characteristics for both calibration and online operation, it ensures a more accurate representation of the secure zone.
- Centimeter-Level Accuracy is Achievable and Practical: The combination of ultrasound positioning and the proposed calibration/zone fitting methods demonstrates that high-precision location-based access control is feasible for practical applications.
- RTJ for Efficient Online Decisions: The RTJ algorithm provides a computationally efficient way to make access control decisions in real-time, which is critical for a responsive system.
Key Results
- The LOCK system can be set up without manual measurement efforts.
- The system can accurately judge physical access authorities with centimeter-level judgment resolution.
- The Maximum False Judge Distance (MFJD) was less than 6cm when using CSZF, compared to less than 14cm when using a manually measured secure table. This highlights the improved accuracy of CSZF.
- Calibration errors for the POD center (x, y) and angle (θ) were within 20cm, 10cm, and 5 degrees respectively, even with a small number of feature points (as few as three), demonstrating the efficiency of MFC.
Limitations and Future Work
- Multiple POD Calibration: The authors suggest that simultaneously calibrating multiple PODs could further improve the ease-of-use of the system.
- Hierarchical Secure Zones: Exploring hierarchical relationships among secure zones could enable automatic generation of location-based secure policies.
- Concave Polygons: While the RTJ algorithm is shown to work for both convex and concave polygons, further exploration of complex secure zone shapes and their implications might be beneficial.
Summary updated