The Study On Channel Estimation Based On Superimposed Pilot For MIMO-OFDM Systems

Multiple input multiple output and orthogonal frequency division multiplexing (MIMO-OFDM) system is the key technique for the next generation wireless communication systems. It can obtain time, frequency and spatial diversity at the same time, and increase the system capacity and the spectrum efficiency. In wireless communication systems, the fading channel has crucial effect on the signal detection and the system design. So it is necessary to study the channel estimation method of MIMO-OFDM systems.There are main two kind of channel estimation methods: pilots-aided channel estimation and blind channel estimation. The channel estimation method based on pilots has received much research. It adopts the algorithms such as least squares (LS), minimum mean square error (MMSE), maximum likelihood (ML), and so on. But this method occupies the special sub-carriers or time slots and then reduces the system spectrum efficiency and data transmission rate. Blind estimation dose not need the pilots and uses the received signal to estimate the channel state information. But the blind estimation algorithm, for example, the sub-space decomposition, is not good for real time estimation because it requires large received signals, and has low estimation convergence rate. The channel estimation based on the superimposed pilot can resolve the problems of spectrum efficiency in the system based on pilot and real time estimation in blind estimation system, for it directly adds the pilots to information symbols without consume extra sub-carriers or time slots. So it is widely studied by researchers. In this thesis, we study the channel estimation methods based on the superimposed pilots.The superimposed pilots have already been applied for channel estimation in OFDM systems and MIMO-OFDM systems, but the existed estimation methods are only suitable for static fading channel or do not consider the disturbance caused by unknown information symbols. Based on the above analyses, we propose the following algorithms:1. We propose the frequency-domain superimposed pilot time-domain correlation (FSPTC) channel estimation method for OFDM system. The channel state information is obtained by the cross-correlation between received signals and the pilots, and the auto-correlation of pilots. But the unknown information symbols disturb the pilots. In order to eliminate the disturbance, we take the Pseudo Random Binary Sequence (PRBS) which has good auto-correlation characteristic as the pilots, and present an iterative estimation method between channel estimation and signal detection. Through theoretic analyses and computer simulations, the effect of the parameters including OFDM sub-carrier number, pilot power ratio and pilot period on estimation mean square error (MSE) and system symbol error rate (SER) are discussed.2. We propose an improved RLS algorithm for superimposed pilot MIMO-OFDM system. The proposed algorithm presents the "decision directed" method to reduce the disturbance caused by unknown information data. Simultaneously, we design the pilots to be the same sequence in all the symbol blocks to decrease the computation complexity.