The Study Of Frequency Offset Estimation Algorithms For OFDM Systems

The network of industry control system makes the use of wireless technique possible. Wireless technique not only researched in theory, but also in some industry controlling system. Orthogonal frequency division multiplexing (OFDM) is a kind of multi-carrier modulation in wireless environments. It is to transfer high-speed serial data streams to low-speed parallel data sub-streams and increase the duration of symbols, which can reduce the ISI caused by multi-path channel.In recent years, as the rapid development of the hardware technique, OFDM would be used in the next wireless communication system because of its high efficiency of using the limited spectrum and its robustness against the effects of multi-path propagation. Wireless network has unexampled advantages, such as, it can excuse a mass of connection in the circuit, save the expenses of foundation and maintenance, settle for some special situations, and improve the flexibility of the system. So, wireless technique has a large space to develop and a big application prospect. While in the meantime, OFDM has its own shortcomings: its sensitivity to a frequency offset. If the OFDM system loses synchronization in frequency, the performance of the system will be bad.How to eliminate the problems is the main task of the research of OFDM. In this thesis, we mainly discuss and put forward the problem of frequency offset estimation.1. Analyze and summarize the influence which frequency offset brings to OFDM systemOFDM technology differs every sub-channel by orthogonality between sub-carries. Because of the time variety, the frequency offset in the signal transmission and the frequency difference between transmitter and receiver both can destroy the orthogonality between OFDM sub-carries, and interrupt the sub-channels. So, the main shortcoming of OFDM system is the sensitivity to frequency offset.2. Improve the performance of the CP method on the basis of time-frequency domain data structureThe method on the basis of time-frequency domain data structure includes the one based on Cyclic Prefix and the MUSIC-like algorithm. The method based on Cyclic Prefix is studied in Gaussian white noise channel and frequency selective fading channels. The simulation conclusion is when time delay of channel is so large that inter-symbol interference presents, the performance declines. Through simulation tests, it is showed that in the environment with small delay spread, when frequency spread of channel varied, the performance of the algorithm is invariable.3. Optimize the performance of the MUSIC-like methodIt is by gradually improving the experiment performance to optimize the parameter of the system and the algorithm. It is showed by simulation tests that the MUSIC-like algorithm has large computational burden. And the MUSIC-like method will degrade considerably in present of ISI even when SNR is high.4. Put forward the Weighting Function algorithm on the basis of the maximum likelihood estimationAlthough the algorithm based on timing domain training sequence outperforms the others, it uses all sub-carriers to transmit training sequence thus consumes much more system resources. Moreover, unless the training sequences are frequently transmitted in system, the algorithm can not be used to track the drifting of the frequency offset and to adapt to fast fading channels. Then, pilot-assisted frequency estimation algorithm based on the use of the maximum likelihood estimate is referred. This method can reach good results in AWGN channels. But it is another thing in multi-path channels, because of the different signal-noise-ratio in different position of the pilot.Then we present a new pilot-assisted frequency offset estimate algorithm based on Weighting Function. In particular, the frequency offsets estimated from pilots in different spectral locations are properly combined together so as to reduce the influence of multi-path channels. Simulations show that the performance of the frequency estimator based on Weighting Function provides an excellent estimate of frequency offset in wireless systems.5. Solve the in-phase problem of OFDM, which can make it possible to control industry equipments exactly by wireless LAN.