A Fault-Tolerant Clustering Approach for Target Tracking in Wireless Sensor Networks

Abstract

Target tracking is a crucial application in wireless sensor networks. Current algorithms for target tracking primarily involve node scheduling based on trajectory prediction. However, when the target is lost due to prediction errors, a target recovery mechanism initiates a search operation, potentially activating numerous nodes and leading to increased energy consumption. Furthermore, the recovery process may result in data loss. To address these challenges, we propose a fault-tolerant clustering approach using the Cat Optimization Algorithm to minimize the probability of target loss. To assess the effectiveness of our approach, simulations were conducted in OPNET using the NODIC, DCRRP, BFOABMS, and AFSRP protocols. The results illustrate that our method excels over existing approaches across various metrics. Specifically, compared to the well-known NODIC method, our approach reduces end-to-end delay by 84.93%, media access delay by 15.08%, increases throughput rate by 3.84%, lowers energy consumption by 4.49%, improves signal-to-noise ratio by 9.99%, and enhances delivery rate of data to the sink by 1.02%. Additionally, compared to the widely recognized DCRRP method, our method improves media access delay by 2.90%, throughput rate by 2.02%, reduces energy consumption by 0.30%, enhances signal-to-noise ratio by 7.36%, and improves the delivery rate of data to the sink by 0.41%. Moreover, our proposed method decreases the end-to-end delay by 10.28% compared to DCRRP. Also, the superior performance of the proposed method in terms of end-to-end delay is 1.52%, media access delay by 8.73%, throughput rate by 1.97%, energy consumption by 0.33%, signal-to-noise ratio by 9.25%, and delivery rate of successfully sending data to the sink is 0.76% higher than the well-known AFSRP method.Additionally, compared to the widely recognized BFOABMS method, our method improves media access delay by 9.56% and enhances the delivery rate of data to the sink by 0.70%. However, in our proposed method, the energy consumption criterion has increased by 13.63%, the end-to-end delay criterion by 50.78%, the signal-to-noise ratio decreased by 15.66%, and the throughput ratio decreased by 26.88% compared to BFOABMS.