Ofdm System Carrier Frequency Synchronization Algorithm

As a special multicarrier modulation technique, OFDM (Orthogonal Frequency Division Multiplexing) has been widely used in data transmission systems such as Digital Video/Audio Broadcasting systems and wireless LAN systems, due to its high spectral efficiency, high data rate, and excellent performance to combat with multipath interference. It is the orthogonality between the subcarriers in OFDM system that guarantees all those strengths mentioned above. However, in practice, the carrier frequency offset, caused by the non-ideal factors of the oscillators and the Doppler Shift introduced by the multipath channel, will destroy the orthogonality between the subcarriers. In order to get a better performance, the receiver should estimate the carrier frequency offset and then compensate the received signal, which is usually called as carrier frequency synchronization or frequency offset estimation for short. OFDM systems are very sensitive to carrier frequency offset, that is to say, a small frequency shift could degrade the system performance dramatically. Therefore, frequency synchronization technique has become one of the key techniques of OFDM systems.The existing frequency synchronization techniques can be divided into two categories: the blind algorithms and the non-blind algorithms. This dissertation first illustrates the basic principles of those two types, introduces some common used algorithms based on training sequence and makes simulations and analysis. Then, two methods employing a threshold or trapezoid approximation are proposed to improve the timing performance of the SC (Schmidl and Cox) algorithm. The new methods can improve the performance of fractional frequency offset estimation by decreasing the timing error caused by the plateau effect. Moreover, according to the cyclic shift property of the received signal caused by the integer frequency offset, this dissertation presents a new integer frequency offset estimation algorithm. After precise timing and fractional frequency offset compensation, FFT is operated on the training sequence. By cyclic shifting the training sequence at the receiver side and detecting the minimum difference between the training sequence at the transmitter side and receiver side, integer frequency offset can be estimated. The estimation range reaches to±[1,N-1] subcarrier spacing, where N is the length of the training sequence. Then, simulations of the improved fractional frequency offset algorithms and the proposed integer frequency offset estimator are accomplished in Matlab to verify the validity of those methods. At the end of this dissertation, the future research directions and problems to be solved are briefly discussed.