| Summary: | 碩士 === 元智大學 === 通訊工程學系 === 95 === OFDM signals may possess high peak to average power ratios (PAPR). While transmitting through an amplifier, high instantaneous power signals may enter the nonlinear region and result in in-band signal distortion and out-band interference. PAPR reduction for OFDM signals is therefore an important research topic. The idea using of linear phase variation to reduction PAPR is to apply linear phase shifts to sub-blocks of sub-carriers in the frequency domain. This is equal to translating (cyclic shifting) partial transmit signal sequences in the time domain. In this thesis, linear phase variation is also called partial transmit sequence translation (PTST).
The concept of using linear phase variation for PAPR reduction was first introduced in [Hsi05] [LH06]. The signal modulation, method of sub-blocks splitting, and performance evaluation results are all limited. Only adjacent subcarrier sub-blocks and PSK-OFDM signal are considered in their simulations. This thesis extends the idea of linear phase variation with PSK-OFDM and QAM-OFDM signals. Methods of grouping subcarriers into sub-blocks include adjacent, interleaved, and pseudo-random. In performance evaluation, PSK-OFDM and QAM-OFDM signals are split into two sub-blocks and four linear phase variations are possible. In each case, when PAPR CCDF equals 10-3, the PAPR can be reduced by about 2dB. Exhaustive tests were also conducted for 8-subcarrier QPSK-OFDM signals split into two sub-blocks and optimized over four linear phase variations. Our results show that the percentage of lowest PAPR sequences selected is 72.27%. Symbol error rate comparison shows that PTST is more robust than PTS. In conclusion, PTST performance the best in all our performs evaluation categories.
|
|---|
