| Summary: | 碩士 === 國立臺灣科技大學 === 電子工程系 === 99 === Since the popularity of wireless online and wireless communications is increasing, the number of users has also increased. For the finite wireless spectrum, we have to divide wireless spectrum effectively and maximize spectrum efficiency. Then we can generate multiple independent parallel channels and transmit data simultaneously by using the multi-input multi-output (MIMO).Therefore using MIMO improves spectrum efficiency without increasing bandwidth. We can obtain the physical layer for 4G wireless communications by combining MIMO with orthogonal frequency-division multiplexing (OFDM).
There are two popular detection algorithms which are discussed widely, Minimum Mean Square Error-Successive Interference Cancellation (MMSE-SIC) and List Sphere Decoder (LSD). In this thesis, we focus on researching LSD in iterative detection and decoding schemes. The objective of my research is to approach the performance of MAP when LSD combined with outer encoder in MIMO-OFDM systems. In multiple-in multiple-out antennas environment, the complexity of MAP detector grows exponentially with the number of transmitter antennas, so we use the Sphere Decoder to reduce the complexity. In MIMO spreading situation, we proposed LSD PLUS NLOGN detection, which is also easier to implement in a practical hardware application. By using iterative methods with LSD, “soft” information were exchanged between detector and decoder, and the performance can be increased.
Finally, in MIMO-OFDM systems, LSD PLUS NLOGN detects the symbol and decodes spreading code efficiently. Moreover, it approaches the performance of MAP efficiently.
|
|---|
