Channel Estimation In Ofdm And Sc-fde Technology

Orthogonal Frequency Division Multiplexing (OFDM) is a multi-carrier technology, which modulates the information onto a group of orthogonal subcarriers, and transmits them in parallel. So. the signal on every subcarrier has a relative longer signal duration, which makes OFDM a robust technology to Inter-Symbol Interference (ISI) caused by multipath propagation. OFDM is an attractive technology for its high immunity to ISI, high spectrum efficiency and low computational complexity, and is considered as an important technology in the Forth Generation (4G) Mobile Networks. Single Carrier with Frequency Domain Equalization (SC-FDE) is another efficient wideband digital communication technology. Besides those advantages that OFDM technology have. its relative lower Peak to Average Power Ratio (PAPR) makes SC-FDE an attractive technology in wideband digital communication. To improve system performance, there are many key technologies need to be solved or improved in OFDM and SC-FDE systems, where channel estimation is one of them.In this thesis, we describe the characteristic of wireless propagation environment and the principle of OFDM and SC-FDE systems in brief. After describing some basic channel estimation algorithms in OFDM and SC-FDE systems, we propose some new channel estimation algorithms.Firstly, by studying the time domain effect of Channel Impulse Response (CIR) and the transmitted OFDM signal, we propose a time domain channel estimation algorithm, which estimates the CIR by solving a bank of equations. In this thesis, we discuss two kinds of training sequence format for this algorithm. For each of them, we seek the optimized training sequence with Generic Algorithm (GA). The results show that, this algorithm with optimized training sequence can achieve good performance, while it's computational complexity and PAPR is low. Furthermore, the training sequence format (b) can be solved in parallel.Secondly, by analyzing the characteristic of SC-FDE signal in time and frequency domain, enlightened by the duality of OFDM signal and SC-FDE signal, we know that SC-FDE signal has a large PAPR in frequency domain. So common data frame can't be used as the training sequence to estimate CIR, and the training sequence must be designed carefully. By simulation, we conclude that Newmann sequence can be used as training sequence.Finally, enlightened by article [49], we can arrange the frequency domain pilot in SC-FDE systems, so we can use those pilot-based frequency domain channel estimation algorithms proposed in OFDM systems to estimate channel in SC-FDE systems. However, simulation results show that both the pilot-based algorithm and the UW-based algorithm proposed in article [43] has low precision. To improve the estimation precision, we propose a feedback-based algorithm, which modifies the coarse estimation result achieved by pilot-based algprithm or UW-based algorithm. Simulation results show that, the feedback-based algorithm can improve estimation precision and decrease the Bit Error Rate (BER) dramatically. The feedback-based algorithm can also be used to track the channel varying by replacing the course estimation result with the CIR of previous frame.At the end of this thesis, we summary the major results obtained of this thesis, and outline possible future work in this field.