The Design And Implement Of Equalization And Channel Estamition In The Up Link Of GSM/EDGE System

As a transitional scheme from General Packet Radio System (GPRS) to third-generation (3G), Enhanced Data rates for GSM Evolution (EDGE) was becoming more and more important for some countries and areas which were in the evolution from second-generation (2G) to 3G because it could reuse current radio resource and equipments. As adopting 8PSK algorithm, the capacity of information of EDGE system was enhanced. However, the number of samples' increasing also made the maximum-likelihood sequence estimation (MLSE) equalizer which was used in GSM system became infeasible. So people began to research the reduced-state sequence estimation.In this paper, we mainly research the design and implement of equalization algorithm in the up link of EDGE system. The paper introduced the theory of channel estimation, pre-filter and equalization, as well as gave the simulation result and the scheme of implement on the hardware.In the part of channel estimation, the paper provided a detailed introduction of joint channel and DC offset estimation. DC offset could not be eliminated completely in the RF part of the receiver and had to be taken care of in the baseband signal processing. The joint channel and DC offset estimation in this paper could reduce the effect of DC offset effectively.In the part of prefilter, the paper introduced the principle of prefilter, the function of prefilter was to change the mix-phase channel into its minimum-phase equivalent which was used to improve the performance of the reduced-state equalization. In the three algorithm introduced in this paper, we finally choose the LP algorithm as the scheme of implement on the hardware because of its least complexity.In the part of equalization, the paper introduced the MLSE algorithm, delayed decision-feedback sequence estimation (DDFSE) and reduced-state sequence estimation (RSSE). The last two algorithms all decreased the number of states to cut down the algorithm complexity by reducing the number of taps. RSSE also used a symbol subset portioning to get a further states reduction.Based of these theories introduced above, this paper also provided a detailed solution based on hardware, which was implement on the DSP chip called picoArray which was produced by picoChip in England.