부산대학교 도서관

Beamforming in distributed antenna array systems requires accurate synchronization of the frequencies and phases of the signals emitted by nodes in the array. However, systematic errors such as oscillator frequency drift and phase jitter cause significant phase differences that decohere the signals if left uncorrected. Furthermore, when synchronizing the frequencies and phases of the nodes, estimation errors can further add to the overall phase error. In this work, we propose a decentralized message passing based average consensus algorithm that synchronizes these parameters across the array to a residual phase error of about 10 −11 degrees in an array consisting of only 20 moderately connected nodes and using an update interval of 0.1 ms. Simulation results are included where the residual phase error and the convergence speed of the MPAC algorithm is compared to the existing decentralized frequency and phase consensus based algorithms by varying the connectivity between the nodes and the number of nodes in the array.