Parallel Search Detection for the MIMO System Based on Differential Metrics

  • 施 惟棠

Student thesis: Master's Thesis


With the development of modern communications large amounts of data are to be transferred rapidly and reliably Therefore the multiple-input multiple output (MIMO) system has become the mainstream of wireless communications because the MIMO system can make full use of spectrum efficiently and increase the transmission throughput As a result the design of low-complexity detection has become a significant issue Although the sphere decoding (SD) algorithm is an efficient approach to obtain the optimum maximum likelihood (ML) detection it still has high complexity especially at high signal-to-noise ratio (SNR) In this thesis we propose an efficient detection algorithm for the MIMO system associated with differential metrics and gradient search First we define the differential metrics and derive the recursive calculation of differential metrics of different orders for gradient search Then we apply the indicative functions to determine some possible ML bits of the initial sequence By simulation we observe the performance of search of different orders and initial sequences We conclude that the maximum order of search dominates the performance We also compare the parallel search with the proposed modified parallel search that uses diverse initial sequences and the latter can not only avoid the complexity of high-order search but also improve the performance
Date of Award2017 Jul 19
Original languageEnglish
SupervisorMing-Xian Chang (Supervisor)

Cite this