November 2025: Congratulations to Alperen Duru for receiving the Best Poster award at the Brooklyn 6G Summit!
6G will have AI applications and optimization as a core component. However, these implementations require dataset collection from the environment, which may be site-specific. Digital twins can help generate synthetic datasets for such AI implementations for deployment in reality, tuned with little environmental data. However, digital twins do not readily exist currently, requiring either generation from scratch or calibration from the available datasets. This study demonstrates two main views of modern digital twins:
(1) How do coarse, but readily-available geometry maps propagate their geometric uncertainties to channel estimation uncertainties?
(2) How can an extended reality (XR), equipped with accurate position & orientation estimates, help calibrate such a map geometry?
