Defending against Routing Attacks with Privacy-Preserved Localization in Wireless Sensor Networks

One of the most significant study scenarios involving security is the development of routing assault localization. Improving security is essential in numerous applications for wireless sensor networks (WSNs). Routing attacks impair network performance by introducing illegal nodes across WSN. This paper uses advanced techniques of Deep learning (DL) on efficient distance measurement and data transmission to offer localization and identification risks of routing through WSNs. Initially, the unknown nodes are detected by using secure localization of Differential Evolution (DE) and Received Signal Strength Indicator (RSSI) methods can potentially pinpoint the locations of unknown nodes even in complex and noisy environments. Then, Malicious attacks are identified by using the Multi-head Self-attention based Bidirectional gated recurrent unit Convolutional neural network (MHSA-BiGRU-CNN) model to identify various types of relationships and features that might be indicative of malicious activity. The Binary Osprey Optimization (BOO) algorithm of routing through the most effective path has been used to improve sensor network efficiency. This optimizationcan help ensure that data packets reach their destination successfully and with minimal loss. The results are analyzed using a MATLAB platform and proposed approach performs significantly higher than the existing approaches.The simulation show that an accurate level of localization for unknown nodes can be achieved with the proposed method. Thus, the proposed model has been measured by various performance metrics, including Average localization error (0.163%), communication radium (0.149%), Energy Consumption in time (4.53 sec), and Network lifetime (5041 sec), respectively.