ASOCIDA: adaptive self-optimizing approach for anomaly detection and collaborative isolation in wireless sensor networks

Ensuring reliable and secure communication in wireless sensor networks (WSNs) is critical for a wide range of applications, including surveillance, energy management, and healthcare systems. However, existing anomaly detection and isolation methods often face difficulties in maintaining high detection accuracy, minimizing false alarms, and conserving energy, particularly under rapidly changing network conditions. This paper addresses these limitations by introducing ASOCIDA (Adaptive Self-Optimizing Approach for Anomaly Detection and Collaborative Isolation in WSNs), a novel and dynamic framework that combines real-time monitoring, collaborative validation, and self-adaptive optimization. The proposed approach leverages Mahalanobis distance and adaptive EWMA for responsive anomaly detection based on dynamic thresholds. A distributed reputation system, enhanced by Byzantine Fault Tolerance mechanisms, is used to validate anomalies collaboratively and ensure robust local isolation. Furthermore, a closed-loop control system dynamically adjusts the detection parameters based on alert reduction trends observed after isolation, ensuring continuous performance improvement. Simulation results demonstrate that ASOCIDA achieves a detection accuracy of 98%, a false alarm rate as low as 0.3%, and a 15% reduction in energy consumption, significantly outperforming traditional techniques. Additionally, it offers an average response time of 15ms and a recovery time of 1.02s, while increasing the packet delivery rate by 10% and reducing the average latency by 20%. These outcomes confirm the potential of ASOCIDA to improve both security and efficiency in dynamic and heterogeneous WSN environments.