A Study On Algorithms Of Channel Estimation For Nc-ofdm Based On Pilot Signal

Today's wireless networks are characterized by a fixed spectrum assignment policy which directly results in the unhealthy usage and scarcity of spectrum resource. To overcome this problem a new communication paradigm referred to as Dynamic Spectrum Access is necessitated which is supported by a novel technology Cognitive Radio(CR for short). NC-OFDM is proposed as one of transmission modes of CR who has all merits of OFDM but as well as many new problems should be studied deeply. One of them is how to design the pilot pattern and how to estimate channel in the transceiver under available subcarriers changing randomly, and this is the core of the paper as well.The paper introduces the principles of NC-OFDM and its characteristics, and then proposes two practical channel estimation methods for IEEE802.16d according to whose physical frame structure and OFDM symbol structure, of which modified transform domain interpolation method based on DFT/IDFT applies to system with uneven comb pilots just as NC-OFDM. For the channel with bigger Doppler frequency shift, a joint channel estimation approach is given. What follows is to stochastically simulate the NC-OFDM system and analyze the characteristics. We model each of chosen subcarriers as continuous-time Markov chain, so whose vacant/occupied state process is assumed to be a Poisson process. Then we design all chosen subcarriers as i.i.d and analyze the characteristics in detail. According to the analyzing results two pilot-inserted strategies respectively based on block pilots and comb pilots are investigated, and the channel estimation approaches mentioned above are applied correspondingly. At the end of the paper simulation results show us the superiorities and feasibilities of the pilot strategies and the channel estimation approaches.