This dataset comprises a collection of synthetically generated wideband signals, which were used in experiments conducted for the paper "An end-to-end deep learning framework for wideband signal recognition," submitted for publication to IEEE Access. In this work, the proposed learning-based approach for signal detection, localization, and classification was evaluated on the public wideband signal recognition dataset introduced by West et al. (https://ieeexplore.ieee.org/document/9593265). We note that the dataset provided here contains only additional auxiliary data that were generated to complement the public benchmark dataset in experiments that employ data mixing and transfer learning techniques. As such, the synthetic wideband signals in our dataset are designed to mimic those in the benchmark dataset. Specifically, in our simulations, a wideband signal is modeled as a superposition of several narrowband emissions and receiver noise. Similar to the benchmark dataset, all wideband signals have a normalized sampling frequency of \(F_\text{s} = 1\) sample per second (sps), duration of \(L = 1 000 000\) samples, and a center frequency of \(F_\text{c} = 0\) Hz. The narrowband signals have varying center frequencies, bandwidths, start times, and durations, and are modulated with a range of modulation classes, including: Amplitude Modulation - Double Sideband (AM-DSB) Amplitude Modulation - Single Sideband (AM-SSB) Frequency Modulation (FM) \(M\)-Frequency-Shift Keying (\(M\)-FSK) for \(M \in \{2, 4\}\) \(M\)-Continuous Phase Frequency-Shift Keying (\(M\)-CPFSK) for \(M \in \{2, 4\}\) Gaussian Minimum Shift Keying (GMSK) On-Off Keying (OOK) \(M\)-Phase-Shift Keying (\(M\)-PSK) for \(M \in \{2, 4, 8\}\) \(M\)-Quadrature Amplitude Modulation (\(M\)-QAM) for \(M \in \{16, 64, 256\}\) The modulation parameters for each narrowband signal are randomly selected from predefined typical ranges. For all wideband signals, the value of the power spectral density of the added white Gaussian noise has been initially set to -174 dBm/Hz. Further noise addition according to the desired signal-to-noise ratio (SNR) level can be applied by the user. The data are stored and documented according to the Signal Metadata Format (SigMF) standard (https://github.com/sigmf/SigMF). Each wideband signal in the dataset is represented by a data file (a binary file containing digital I/Q samples of the signal) and a metadata file (in JSON format) that annotates the signal. The annotations include the general properties of the wideband recording (sampling rate, center frequency, signal duration, and noise power spectral density), as well as specific information for each narrowband signal (start time, duration, lowest and highest frequency, modulation class label, and power). For every signal in the dataset, the data file and the annotations file share the same name. For example, the first training wideband signal is stored in the file "train_1.sigmf-data" and is annotated by the file "train_1.sigmf-meta".

{"references": ["N. West, T. O'Shea and T. Roy, \"A Wideband Signal Recognition Dataset,\" 2021 IEEE 22nd International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Lucca, Italy, 2021, pp. 6-10, doi: 10.1109/SPAWC51858.2021.9593265.", "B. Hilburn, N. West, T. O'Shea, and T. Roy, \"Sigmf: the signal metadata format,\" in Proceedings of GNU Radio Conference, vol. 3, no. 1, 2018."]}