Our new and re-processed datasets provide additional Doppler-domain synchronization guarantees.

Use insights into the physics of wave propagation to learn a highly accurate map of the radio environment in a self-supervised manner.

Leverage a self-supervised triplet learning technique to create a virtual map of the radio environment from radio channel observations and timestamps.

Use the uplink channel state information available at the base station to predict the downlink channel at a different frequency band.

Train a neural network to perform angle of arrival estimation and see how it generalizes to unseen data.

Compensate the effects of REFTX channel phase shifts and propagation delays using the provided values.

Train a TensorFlow-based deep neural network to perform indoor transmitter positioning in a well-known physical environment.