CRAWDAD OWL-InIT

Traces of various radio signal standards  Dataset of traces of IEEE 802.11b/g, IEEE 802.15.4 and Bluetooth packet transmissions with varying SNRs in the baseband. Note: Python 2.7 is required to load the dataset. Additionally, different frequency offsets were added in the baseband to reflect different channels of the wireless technologies. date/time of measurement start: 2016-08-22 date/time of measurement end: 2016-08-22 collection environment: We have recorded three different wireless technologies. The goal was to create a data set in order to train and validate artificial neural networks which can classify wireless interferences.network configuration: The measurement recordings are intentionally limited in their complexity and variation, with the aim of designing a first well-functioning prototype of a pattern recognition system. As the number of traces increases, the complexity of the measurement recordings also increases. The detailed network configuration is described in the individual traces.data collection methodology: The data was generated with the vector signal generator (VSG) SMBV100A by Rohde and Schwarz and recorded with Tektronix's RSA6114A real-time spectrum analyzer (RSA). The payload was random. Varying SNRs and a frequency shift in the baseband were added in a SIMULINK model afterwards.limitation: For different frequency offsets the same records were used, meaning that the symbol stream repeats for different frequency offsets.Tracesetemissions-single-label Measurements of IEEE 802.11b/g, IEEE 802.15.4 and Bluetooth emissions file: owl-interference.tar.gzdescription: In order to minimize channel influences, we have connected transmitter and receiver with a coaxial cable. At the same time, only one wireless technology is active. Later, a frequency offset was added to reflect different channels of the wireless technologies. Additionally, white noise was added to the data during post processing.measurement purpose: Network Diagnosis, Usage Characterization, Network Security, Opportunistic Connectivitymethodology: The measurements are recorded with a real-time spectrum analyzer (RTSA) with a sampling rate of 50 MHz and receive filter bandwidth of 40 MHz. Then, they were resampled to a sampling rate of 10 MHz. The transmission power of the vector signal generator (VSG) has been set to -30dBm. The final resulting start and stop frequencies are 2421.5 MHz and 2431.5 MHz (center frequency of 2426.5 MHz). RSA and VSG used a synchronized oscillator.emissions-single-label TraceM1: Single label data with three wireless technologiesfile: WirelessInterferenceIdentificationDataset.tar.zipformat: The data were stored in pickle files. The data structure is a 5-dimensional tensor of the following structure: 15 x 21 x 715 x 128 x 2 - tensor Label x SNR x Snapshot x Samples x iq Label: 0..9: Bluetooth channel with center frequency of 2422 MHz Label * 1 MHz 10..12: IEEE802.11b/g channel with center frequency of 2422 MHz (Label-10) * 5 MHz 13..14: IEEE802.15.4 channel with center frequency of 2425 MHz (Label-13) * 5 MHz SNR: 0..20: SNR between -20 dB...20 dB with the step size of 2 dB, (e.g. index 6 referres to SNR of -8 dB). Snapshot: 0..714: Number of the snapshots per Label Samples: 0..127: Sample number per snapshot iq: 0: In-Phase 1: Quadrature The data can be unpickled in Python with the following lines: import cPickle with open(path_to_file, 'rb') as fo: data = cPickle.load(fo)