Research On The Algorithm For Carrier Synchronization In MIMO-OFDM Systems

With the development of the communication technology, the people's demond for the capacity and bandwidth of the wireless communication systems is growing. So during the research process of the 4th mobile communication technology, the MIMO-OFDM technology which has a large capacity and high spectrum efficiency has becomed a new research focus. By setting a number of antennas both in the receiver and the sender, the MIMO technology can obtain a high capacity without the expanding of the bandwidth. The OFDM technology can obtain high spectrum efficiency by using a number of orthogonal sub-carriers to send data. And the MIMO-OFDM technology combines their virtues.Because the sub-carriers need to be strictly orthogonal, the OFDM technology is very sensitive to the carrier frequency offset. And this disadvantage is inherited by the MIMO-OFDM technology. Furthermore, because of the multi-input and multi-output structure, the issue of carrier synchronization has become more complicated.In this paper, in order to clearly elaborate the issue of carrier synchronization in MIMO-OFDM systems, we introduce the OFDM technology, MIMO technology and MIMO-OFDM technology at first. Then we analyze the influence of the carrier synchronization error to the system performance. And we also introduce several classical algorithms both for the OFDM systems and the MIMO-OFDM systems. At last, based on the reading of the literatures and the research, we propose two new algorithms for carrier synchronization.Nowadays, the algorithms for carrier synchronization in the MIMO-OFDM systems are generally based on the design idea of training sequence. There are two main problems in the design of this kind of algorithms:First, how to distinguish the signals from different transmitting antennas. Second, how to expand the estimation range of the carrier frequency offset. For the solution of the first problem, the two algorithms that we proposed in this paper both use the orthogonal code to design the training sequence. For the solution of the second problem, the first algorithm we proposed use a special training sequence which is composed by many repeated parts. And the more parts we send, the larger range we obtain. The second algorithm we proposed still use a traditional kind of training sequence to obtain a large estimation range. This kind of training sequence is composed by two identical halves. Using this structure, we can estimate the fractional frequency offset. Meanwhile, for the estimation of the integral frequency offset, we design a new estimation algorithm that we can use the differential coding in the internal of the sequence to estimate the integral frequency offset.In this paper, the simulation is carried out for both of the two algorithms. And according to the simulation results, we analyze the estimation performance, advantages and disadvantages of the two synchronization algorithms.