Localization of Underwater Wireless Sensor Networks for Ranging Interference based on the AdaDelta Gradient Descent Algorithm

In Underwater Wireless Sensor Networks, ensuring precise localization of sensor nodes is crucial. Existing localization algorithms heavily rely on pre-deployed and known anchor nodes to compute and determine the positions of other nodes. However, anchor nodes may be compromised or misled in practical scenarios, resulting in localization inaccuracies. To address this challenge, a proficient iterative localization algorithm is proposed in Garg and Varna (IEEE Trans Inf Forensics Secur 7: 717–730, 2012), which selects known nodes as reference anchors to locate unknown nodes and analyzes their performance under non-collaborative attacks. Therefore, a novel localization algorithm is introduced to counter collaborative attacks, utilizing time difference of arrival ranging mapping and selective minimum gradient principles from the AdaDelta Gradient Descent (AGD) algorithm to iteratively eliminate deceptive information transmitted by interfering nodes, thereby enhancing localization accuracy. Validations demonstrate its efficacy in combating collaborative network attacks under certain conditions and significantly improving localization performance. Simulation experiments confirm its robustness even under network attacks, ensuring the reliable operation of UWSNs.