LocUNet is a deep learning method to localize a UE based solely on the reported signal strengths from a set of BSs. The method utilizes accurate estimates of the radio maps of each BS, and is tailored to work in realistic propagation environments in urban settings.

See the paper Real-time Outdoor Localization Using Radio Maps: A Deep Learning Approach for more information.

The associated RadioLocSeer dataset is available at the IEEE Dataport and on the project main page. Details of the dataset can be found in our dataset paper Dataset of pathloss and ToA radio maps with localization application.

For reproducibility, see the compute capsule.