Аннотация
Continuous object tracking in IoT-enabled Wireless Sensor Networks (WSNs) requires fast, energy-efficient boundary detection, especially for dynamic phenomena such as toxic gas leakage or wildfire spread. However, existing approaches often rely on cloud-centric processing, resulting in high transmission delays and excessive energy consumption due to large-scale node activation. This paper proposes Boundary Detection of Continuous Objects (BDCO), a fog-assisted scheme that reduces communication overhead and improves boundary accuracy. BDCO organizes the network into grid-based clusters where the Cluster Head (CH) filters anomalous data using a selective aggregation mechanism and forwards only relevant boundary-related information to the fog node (FN). The FN then applies a convex-hull-based boundary estimation model, enabling precise boundary formulation while minimizing node activation. The proposed scheme is implemented in NS-2.35 and demonstrates substantial improvements in energy consumption (3.00E+06), service delay (22 ms), end-to-end delay (33 ms), packet loss ratio (3.0), and boundary accuracy (0.85) compared to existing approaches. Overall, the BDCO scheme provides a more energy-efficient and delay-aware solution for real-time boundary detection of continuous objects in an IoT-enabled WSN environment.