Research Of Channel Estimation In MIMO-OFDM System

The transfer of the information has become one of the most important signs of modern society.Wireless mobile communication has become the necessary communication method in information society.Because of people's more and more desire on all kinds of data service, it has been the main target of the next generation communication system to support high speed multi-media data service.From the development of the service and technology: On the one hand,as the development of the service,the wireless resource has become more and more limited.How to use the limited communication resource becomes the focus of the new wireless communication technology.On the other hand,because the computer and digital signal processing technology's been more mature,communication terminals are able to deal with signals in space domain,in time domain and in frequency domain.The technology in existence - OFDM and MIMO interest people most.MIMO and OFDM are both hotspots in the mobile communication research field, and also are the key technology of next generation communication system.So the research of MIMO-OFDM system has been a great subject in the communication system.OFDM(Orthogonal Frequency Division Multiplexing) transforms the frequency selective channel to the flat channel in frequency domain, decrease the impact of the multipath fading.It has lots of advantages,such as supporting high speed transmission, confronting the multipath fading and efficient spectrum using rate.MIMO(Multiple-Input Multiple-Output) is able to use the multipath effect efficiently in the wireless channel.It generates separated parallel channels in the space,and transmits many pathway data stream in the same time.In this way,it increases the transmission speed.So it increases the capacity of the system and the spectrum using rate without changing the system bandwidth.So if we join MIMO and OFDM together to promote the system's capacity and its performance, it will be a challenging research.And also it will be the key technology in the next generation mobile communication system.In order to make a good performance,the sink of MIMO-OFDM system needs to know the channel state information (CSI).Then it can make a correlated demodulation.So that channel estimation is one of the most important technologies in MIMO-OFDM system.My paper is mainly about the channel estimation algorithms in MIMO-OFDM system.My research begins with basic theories.First, We build the system model based one the mastering the basic principle of MIMO-OFDM system.Second, We research the channel estimation algorithms in MIMO-OFDM system and give the improved algorithm.Last, We analyze the performance of the algorithms through simulation.There are three kinds of channel estimation methods:1. Channel estimation algorithm based on pilots.2. Channel estimation based on decision making feedback.3.Blind channel estimation.The first and the third are our emphases in the paper.Firstly,in the Channel estimation algorithms based on pilots.We have a research on the LS channel estimation algorithm in MIMO-OFDM system.The structure of LS algorithm is simple,and the amount of its calculation is small.But it is sensitive to the AWGN and ICI.When the value of AWGN and ICI becomes big,the performance of LS algorithm will drop.LMMSE algorithm smooths result of LS.So it has a good performance.But you have to know the channel statistical feature.Its calculation amount is big,and it is not fit for the actual application. In allusion to the disadvantages of the algorithms above,We give the improved LS algorithms based on transform domain filtering.We transform the frequency response of LS algorithms into transform domain by FFT.Then we pass it to a low pass filter.After that we transform it back to the frequency domain.During the course,we improve the design of the low pass filter and make its end frequency change according to the SNR in a certain range.We prove the new method really works through simulation.Our improved algorithm make a good performance under a low complexity.Secondly, In the Blind channel estimation algorithm.We use the no memory feature of subcarrier channels and correlation of subcarrier channels nearby,we give a new algorithm in MIMO-OFDM system.We divide N subcarrier channels into M groups, and make sure that the channels are integrated.Because of the correlation of subcarrier channels,We treat thses subchannels as the same ones.In order that we just need to pick up one channel in a group and make the channel estimation.So the complexity is M/N .At the same time,there is a problem in blind channel estimation.That is the amount of sample data is very big.The sink has to receive a lot of data to make one estimation.Once again,we use the correlation of subcarrier channels,we adds the data together,we treat the data received in the whole group as the channel which is selected to make the estimation its own.So that the sink is able to make an estimation by receiving only a few data.It saves time,and it increases the estimation speed.In the MIMO-OFDM system using our new algorithm,every subcarrier channel which is selected is a no meory channel.There are lots of no meory channel estimation methods.We use OPDA algorithm based on only second order statistic.OPDA algorithm estimates H_a H_a~H by second order statistic.Then it will make a matrix HH~H.That is a effective method and not too complex.Lastly,we prove the performance of the blind algorithm by simulation.