Research On Pilot-based Channel Estimation Technologies For MIMO-OFDM Wireless Communication System

With the rapid growth of the demand for mobile communication services in modern society,the spectrum resources become increasingly tense.Therefore,in order to implement reliable and high-speed wireless communication in the future,it is necessary to achieve communication systems with high capacity,high rate and high spectrum utilization.For the advantages of high spectral efficiency,low complexity and anti-multipath fading,OFDM technology is widely used in broadband communication.At the same time,without occupying additional spectrum resources and increasing the transmission power,MIMO technology makes full use of space resources,effectively enhance the system's channel capacity and spectrum utilization.Therefore,it is conducive to the promotion of system capacity and spectrum utilization with the combination of MIMO and OFDM.In the fourth generation of mobile communication system,MIMO-OFDM has been selected as the core technology of data transmission and greatly enhances the users' communication experience.However,the channel transmission environment is complex in the MIMO-OFDM wireless communication system,at the same time,it is essential for space-time decoding,coherent demodulation and restoration of initial data to acquire accurate channel state information.Therefore,channel estimation is the key to reliable communication system and the algorithms of channel estimation in MIMO-OFDM wireless communication system is the main research in this paper.In this paper,the frequency selective fading and non-frequency selective fading MIMO channel models are established according to the statistical characteristics and fading characteristics of wireless channels.Then,the data transmission model of the MIMO-OFDM system is established based on the characteristics of parallel transmission signals through multi-carrier of OFDM technology.After the establishment of system model,three classical channel estimation algorithms,LS,DFT and MMSE are deduced and simulated.In the section of algorithm improvement,an improved DFT algorithm based on cyclic prefix is proposed in order to solve the shortcomings of noise existing in the maximum delay length and the acquisition of the maximum delay length.The maximum delay length is replaced with the cyclic prefix length,and a threshold is set to obtain the main path of the channel within the cyclic prefix length,thus improving the accuracy of the classical DFT algorithm.At the same time,in view of the disadvantage of noise existing in each channel path,the exponential smoothing based improving method for LS and DFT algorithms are proposed to optimize the estimation accuracy.Theoretical analysis and simulation results show that the cyclic prefix based improving DFT algorithm improves the precision of classical DFT algorithm and avoids the problem of obtaining the length of maximum delay.In the improving method with employment of exponential smoothing algorithm,the estimation precision of the classical LS and DFT algorithms are improved as the noise on every path of the channel is reduced.Moreover,the computational complexity of the improved method is relatively low and it is easy to implement.