Beamforming Training Using Common Sequences for Millimeter-Wave Systems

The patent segments the beamforming training process into distinct phases: downlink training where user equipment determines analog weight parameters based on common sequences, and uplink training where the base station determines analog weight parameters based on pilot signals. This segmentation allows complexity to be distributed and managed separately at different nodes, reducing overall system complexity while maintaining optimal precoding performance through specialized processing at each stage.

The patent applies preliminary action by having user equipment determine analog weight parameters in advance during downlink training based on common sequences before actual data transmission. This preliminary determination of weight parameters eliminates the need for exhaustive search during data transmission, as the optimal parameters are pre-established through initial training phases, thereby reducing computational complexity while maintaining selection accuracy.

The patent merges the downlink and uplink training processes into a unified beamforming training framework where both user equipment and base station determine analog weight parameters through coordinated training sequences. This merging eliminates the need for multiple separate feedback transmissions required in multi-layer feedback mechanisms, as the training is accomplished through direct parameter determination at both ends using training sequences, thereby reducing resource consumption while maintaining manageable complexity.