Mixed-Signal Beam-Tracking and Precoding Algorithm with Partial Update and Neighboring Search for Millimeter Wave MIMO Systems

• Main Authors: Yeh,Che Chuan, 葉哲銓
• Other Authors: Huang,Yuan Hao
• Format: Others
• Language: en_US
• Published: 2015
• Online Access: http://ndltd.ncl.edu.tw/handle/64381748571215243791

碩士 === 國立清華大學 === 通訊工程研究所 === 104 === The precoding and large scale antennas system are important technology in millimeter wave for compensating the severe path loss [1–3]. However, the several high cost and high power consumption of the analog-to-digital converters (ADC) in gigabit per second are inevitable if the precoder and combiner are all implemented on baseband, because each antenna should has one RF chain and ADC.Therefore, the hybrid precoding is proposed to reduce the number of RF chain and ADC in [4–6] owing to the limited scatterings in mmWave. Furthermore, the low complexity mixed-signal precoding algorithm is proposed in [7] to reduce the complexity by simplifying the method to find out the precoder in RF chain with the negligible performance degradation. However, despite applying the low complexity algorithm, the computation complexity in searching the bases vector is still high. The propose algorithm in this study can reduce the expected value of computation complexity by keeping the unchanged precoder in RF chain and partially updating the precoder in baseband owing to the high correlation between the neighboring channel state information. The two update condition are also proposed to make the condition to do the update and one searching algorithm is proposed to limit the search range if the AoA and AoD change continually. In final, the simulation results show that the expected value of computation complexity with proposed update condition and proposed searching algorithm simulated in the extended SV channel model with channel coefficient in Quadriga can reduce 27.0%∼31.7% in precoder and 52.2%∼58.1% in combiner according to the different tracks or different velocities. Furthermore, theproposed searching algorithm can reduce 38.3%∼62.4% in precoder and 34.0%∼58.1% in combiner also indifferent tracks or different velocities.