Node identification based on unique hardware features like clock skews has been considered an efficient technique in wireless sensor networks (WSNs). Spoofing attacks imitating unique hardware features, however, could significantly impair or break down conventional clock-skew-based node identification due to exposed clock information through broadcasting. To defend against Spoofing attacks, we propose a new node identification scheme called node identification against Spoofing attack (NISA). It utilizes the reverse time synchronization framework, where sensor nodes’ clock skews are estimated at the head of a WSN, and the spatially-correlated radio link information to achieve simultaneous node identification and attack detection. We further provide centralized and distributed NISA for covering both single-hop and multi-hop scenarios, the former of which employs a single-input and multiple-output convolutional neural network. With a real WSN testbed consisting of TelosB sensor nodes running TinyOS, we investigate the identifiability of clock skews under temperature and voltage variations and evaluate the performance of both centralized and distributed NISA. Experimental results demonstrate that both centralized and distributed NISA could provide accurate node identification and Spoofing attack detection.

中文翻译：

---

NISA：基于时钟特征和无线电信息的无线传感器网络节点识别和欺骗攻击检测

基于时钟偏差等独特硬件特征的节点识别被认为是无线传感器网络 (WSN) 中的一种有效技术。然而，由于通过广播暴露时钟信息，模仿独特硬件功能的欺骗攻击可能会显着削弱或破坏传统的基于时钟偏差的节点识别。为了防御欺骗攻击，我们提出了一种新的节点识别方案，称为针对欺骗攻击（NISA）的节点识别。它利用反向时间同步框架，其中传感器节点的时钟偏差在 WSN 的头部估计，以及空间相关的无线电链路信息来实现节点识别和攻击检测的同步。我们进一步提供集中式和分布式 NISA 来覆盖单跳和多跳场景，前者采用单输入多输出卷积神经网络。通过由运行 TinyOS 的 TelosB 传感器节点组成的真实 WSN 测试平台，我们研究了温度和电压变化下时钟偏差的可识别性，并评估了集中式和分布式 NISA 的性能。