Research On Papr Problem In OFDM System

Orthogonal frequency division multiplexing (OFDM) is a special kind of multi-carrier transmission scheme, which can be viewed as either a modulation technique or a multiplexing method. Because its advantages such as high spectrum efficiency, can effectively confront ISI and low cost, OFDM scheme has been widely used in Asymmetric Digital Subscriber Loop (ADSL), Digital Audio Broadcasting (DAB), High-Definition Television (HDTV), Wireless Local Area Network (WLAN) and other systems. However, in OFDM system the transmitted signals can bring a great peak to average power ratio, so that the linearity of the amplifier within the transmitter is imposed very high demand. In order to solve this problem, many domestic and foreign scholars have done a lot of work on PAPR reduction technology.In this thesis, firstly, the basic principle of the OFDM system is analyzed deeply. The definitions of peak to average power ratio and its probability distribution are introduced, and a comprehensive analysis for the reason resulting in high peak to average power ratio is given as well. On this basis, various technologies of PAPR reduction are introduced, including signal pre-distortion, encoding and signal scrambling. Next, Companding Transform of signal pre-distortion technology, Selected Mapping (SLM) and Partial Transmit Sequence (PTS) of signal scrambling technology are discussed in details.Due to its simple implementation, Companding Transform technique has attracted a lot attention, but will reduce the reliability of the system. For this deficiency, this thesis introduces a Linear Companding Transform (LCT) which not only decreases the bit error rate but also improves the PAPR performance when compared to the traditional companding algorithm based onμ-law.Selective Mapping (SLM) program normally has high computational complexity. In this thesis, using the relationship between non-periodic auto-correlation function of OFDM symbols and PAPR, we employ Gold sequences to improve the non-periodic auto-correlation function of transmission symbols. Simulation results show that Gold sequence as phase sequences of SLM can effectively reduce the system's computational complexity and requires less side information, moreover, can improve the system's bandwidth efficiency.For the Partial Transmit Sequence (PTS), the largest defect is still great computational complexity. This thesis discusses three algorithms which are the Cimini Iteration, Sub-block Iteration and Amplitude Abatement, and compares their computational complexity and PAPR performance. Simulation datas show that Sub-block Iteration can acquire a good compromise between computational complexity and PAPR performance, the method of Amplitude Abatement can significantly reduce the complexity of PTS and improve the PAPR performance betterly.Finally, the combination improvement program of PTS and Linear Companding Transform is proposed. This method combining the combination of PTS and Non-Linear Companding Transform based onμ-law has better PAPR performance and lower bit error rate.