Study On The Application Of GDFT In OFDM Wireless Communication System

In contrast with DFT, there are infinitely many possible GDFT sets with constant modulus available due to the introduction of nonlinear phase. The additional flexibility of changing phase functions of GDFT set allows us to exploit the phase to design various GDFT’s where OFDM performances in a multicarrier communications system might be improved over the existing solutions like DFT.Thanks to its strengths, such as high spectral efficiency, strong ability to resist Inter-symbol Interference (ISI) and high-speed data transmission, OFDM technology has been favored by researchers and becomes the core physical layer technology for next generation of mobile communication system. However OFDM system has its faults, such as sensitivity of frequency offset, high Peak-to-Average Power Ratio (PAPR) and phase noise.Based on the research of GDFT principle, this paper focused on the application of GDFT in OFDM system. Specifically we discussed how to effectively reduce the OFDM signal PAPR value and obtain more effective symbol synchronization method with the help of GDFT.For the application in PAPR problem, we first discussed the existing classical algorithm of reducing the PAPR. The advantages and disadvantages of each algorithm were compared. Then based on the GDFT transform theory, we improved GDFT-OFDM algorithm. By changing the phase of transform phase function, we got the transformed output signal with minimum PAPR. After simulation, the algorithm we adopted is superior to traditional SLM-OFDM and PTS-OFDM algorithm.For the application in symbol synchronization, we first analysised the exiting classical symbol synchronization algorithms such as non-blind symbol synchronization and blind symbol synchronization. Specifically, we discussed the classical algorithm Schimidl&Cox in detail. The platform area of Schimidl & Cox which results in clock synchronization ambiguity is one of its drawbacks. Then we focused on how to improve the performance of Schimidl&Cox. First, we proposed an improved frame structure algorithm so that we can remove the platform area and got precise symbol synchronized location. After that we used IGDFT to generate the first training sequence which has better correlation properties. After simulation, the new algorithm’s noise immunity is significantly enhanced.