Research Of Pilot-based Channel Estimation In OFDM System

With the requirement for high-quality wireless multimedia services, the future mobile communication systems (LTE/4G) require for faster transferring rate, better traffic quality and more spectrum efficiency. Orthogonal Frequency Division Multiplexing (OFDM) technology is widely used in future mobile communication systems more and more because of its advantages for broadband communications. However, due to the transmission characteristics of wireless channel and high-bandwidth nature of OFDM system itself channel estimation is a very important part in OFDM system which determining the system performance.The paper first briefly introduces the characteristics of wireless communication system, and then leads to one of the key technologies for the future mobile communications-OFDM, classified and compared the existing channel estimation methods for OFDM system. These algorithms subjected to Nyquist criteria in the case of uniform distribution of pilot data, thereby reducing the pilot distribution flexibility. In this paper, a new channel estimation method based on the non-conventional pilots is studied, in which channel state information is estimated indirectly using of non-uniform sampling technique, and permit a flexible distribution of pilot data. The simulation results show that non-conventional pilot method can estimate channel parameters well even in fast fading channel conditions. For slow-varying channel conditions the estimation algorithm is improved, simulation shows that the computation can be reduced up to 50%.OFDMA is a good access technology for multi-user broadband wireless communication system, as a result of irregular distribution of user data, the traditional method of channel estimation is very difficult to apply in the OFDMA system, or even can it needs a lot of pilot data; non-conventional channel estimation method in our paper is suitable for the system. The simulation results show that with the larger Doppler shift, non-conventional method has better performance compared with the least-squares (LS) estimation method. When more time-frequency resources are assigned to the user and pilot data remains, the simulation results show that the channel state information can be estimated well, has more transmission efficiency.