Computer Networks ( IF 5.6 ) Pub Date : 2020-09-23 , DOI: 10.1016/j.comnet.2020.107566 Amani Al-shawabka , Francesco Restuccia , Salvatore D’Oro , Tommaso Melodia
Recent research has proved the effectiveness of neural networks (NNs) in “fingerprinting” (i.e., identifying) wireless radios, by determining the hardware impairments emitted from the transmitter during the waveform transmission process. The artificial neurons of the NN layers are employed to identify and track the radios’ unique impairments by training a large amount of raw data released from these radios. Today, the radio fingerprinting field lacks such a large-scale waveform database that can provide a standard benchmark for researchers working on this field. In this paper, we publicly share 2TB of IEEE 802.11 a/g (WiFi) data obtained from 20 bit-similar Software-Defined-Radios (SDRs).
中文翻译:
用于WiFi无线电指纹识别的大规模I / Q数据集
最近的研究已经证明了神经网络(NNs)通过确定在波形传输过程中从发射器发出的硬件损伤,在“指纹”(即识别)无线电中的有效性。NN层的人工神经元通过训练从无线电广播中释放的大量原始数据来识别和跟踪无线电广播的独特损伤。如今,无线电指纹识别领域缺乏如此庞大的波形数据库,无法为从事该领域研究的人员提供标准基准。在本文中,我们公开共享从20位相似的软件定义无线电(SDR)获得的2TB IEEE 802.11 a / g(WiFi)数据。