AI-based Localization: Learning from Synthetic Data from a Genetically Guided Digital Twin

Loading...
Thumbnail Image
Author
Beysens, Jona
Narduzzi, Simon
Bergamini, Lorenzo
Abstract
Synthetic datasets for training a machine learning model is a promising approach to reduce the need for real data. However, the simulated data should be representative of the real data. As a consequence, the tuning of the simulation parameters is critical to reproduce the real environment in the most accurate way. At CSEM we design and evaluate a genetic algorithm (GA) that proposes the optimal parameters for the network simulator.
Publication Reference
CSEM Scientific and Technical Report 2022, p. 24
Year
2022
Sponsors