Research On Technique Of Fast Channel Estimation

This thesis mainly discusses the techniques of channel estimation in Frequency Domain, which utilizes the optimized leading sequence. An enhanced method for searching optimized leading sequences is introduced based on existing documents. This method can decrease the errors and make the searching and caculating easier. The algorithm was validated with simulation.Firstly, the author introduces the criteria of searching leading sequences. Due to the needs for the flat spectrum character, the leading sequence requires high self-correlation. The author selected a series of sequences whose highest peek to second highest peek ratio is big, from a series of 'm' sequences. Then, the optimized leading sequences were judged according to GLF, the sequences which have lower GLF would be more proper for the estimation. After that, the author applied FFT techniques on the real-time channel estimation simulation and it is finally implemented with DSP techniques. From the simulation, the author discovered that channel parameters can really be retrieved rapidly while estimating channel in Frequency Domain, and the caculating work for DSP is not complicated. While working on this thesis, the author also performed channel estimation in Time Domain with some common rapid estimation algorithm to make comparison. The result shows that these two methods can both achieve the rapid estimation.Although this thesis only involves short wave channels, the proposal can also be applied for mobile channels and wireless channels, and so on, and it is valuable for applications.