A hybrid classical-quantum autoencoder system compresses and reconstructs CSI feedback from clustered UEs using quantum annealing for L1-norm optimization.

Patentradars sammanfattning

This invention presents a method for efficient Channel State Information (CSI) compression in wireless networks by clustering User Equipments (UEs) and training a hybrid classical-quantum autoencoder for each cluster. The classical encoder is deployed to UEs for local CSI compression via random projections, while a quantum annealer at the network side reconstructs CSI by solving an L1-norm minimization problem. The autoencoder is periodically fine-tuned at predetermined events, and a quantum encoder unitary is trained using encoded data passed alongside a reference state to improve reconstruction fidelity.

AI-genererad och redaktionellt processad.