Abstract
Bluetooth low energy (BLE) beacons have driven the development of emerging proximity-based services (PBS), which are context-aware applications delivered subject to Proximity of Interest (PoI). Considering the massive beacons deployment in a vicinity, it poses a challenge to identify the correct PoI when the density of a beacon network (i.e., the number of beacons associated to multiple PoIs belonging to the same region) increases. Most commercial applications use a sequential proximity detection approach with fixed scanning mechanism to decide its PoI. Such sequential execution, even though is able to produce reliable detection, it surfers from severe deterioration in connection with detection accuracy when the beacon network gets denser. To address the proximity detection issue in a dense beacon network, an empirical investigation was conducted to study the statistical properties of both receiver signal strength (RSS) and signal interarrival time. With reference to the statistical insights acquired from empirical analysis, this paper proposes a high resolution proximity detection approach using adaptive scanning mechanism fusion with spontaneous Differential Evolution (AS+sDE). Through this novel approach, the receiver is able to adapt its scanning conditioned on network density and make spontaneous decision in parallel with the scanning process. Two performance measures, i.e., detection accuracy and accuracy rate, are used to benchmark the proximity detection performance. The results obtained from both simulations and real-world implementations prove the feasibility of the proposed approach. Further, AS+sDE achieves a superior performance with a detection accuracy of more than 90% in both horizontally and vertically aligned beacon network, and a high accuracy rate, i.e., average < 1s is spent to achieve 90% accuracy.
Paper(s)