UWSig-26: A Synthetic Urban Wireless Signal Dataset for Joint Wireless Technology and Modulation Recognition

UWSig-26 is a synthetic dataset designed for joint wireless technology and modulation recognition in localized urban radio environments. It contains 10 wireless-technology classes, 10 modulation classes, and 28 valid technology–modulation label pairs derived from 32 enabled waveform profiles. All signals are represented at a common sampling rate of 20 MHz using 4,096-sample complex I/Q windows. The dataset covers three representative scenarios (OfficeLab, SmartFactory, and CampusOutdoor), five reception conditions (clean, AWGN, clock, fading, and composite), and signal-to-noise ratios from -20 to 18 dB. The release provides HDF5 data, fixed data splits, metadata, validation materials, and companion generation and audit code. Channel conditions follow controlled, 3GPP-inspired parameterization and are not intended as a complete implementation of any communication standard or full 3GPP spatial channel model.