Study On The Key Technology Of Blind Estimation And Identification For Ofdm Signals

OFDM (Orthogonal Frequency Division Multiplexing) technology is widely used in broadband wireless communication systemthe of civilian area. In order to monitor the civil communications signals for the government departments, achieve electromagnetic interference identification and spectrum management, prevent the illegal use of radio spectrum and interference, and ensure the normal legal communications, it's need for OFDM signal recognition. The applications of OFDM technology are also more and more widely in military field, in order to monitor the activities of the electromagnetic spectrum in the battlefield, achieve threat identification, help to select the electronic interference strategy,even intercept the enemy's useful information, it also needs to identify the OFDM signal. On the other hand, with the development of communications technology to wireless and broadband, spectrum resources are increasingly shortage, however, the allocated frequency resources are idle in time and space with some degree, In order to reasonably use the scarce spectrum resources, cognitive radio technology is one of the key technologies to solve this problem. OFDM technology not only has a high spectral efficiency, but also meet the modulation needs of cognitive radio, therefore it will be used extensively for cognitive radio systems. The key technologies of cognitive OFDM system is to solve the spectrum sensing and signal recognition problems of OFDM signal for sub-user. Existing research focus on the long cyclic prefix OFDM signal, identify the OFDM signal with a correlation algorithm in ideal Gaussian channel conditions, and the recognition rate is not high under the condition of low SNR, it can not meet the needs of practical applications. To speed up the theory research of OFDM signal recognition system, develope the key equipment and systems with independent intellectual property rights, will undoubtedly enhance the national hi-tech innovation and promote the construction of military information, improve the national comprehensive strength and competitiveness further. Some research work is following:First, the background of the research is introduced, pointed out the meaningful questions to identify OFDM signals by estimating the sub-carrier spacing, and then on can rapidly identify the sub-user system. After analysising the main limitations of existing technology, we need the new reliable methods. the importance of DOA parameter estimation for OFDM signal is also discussed in detail. Second, All kinds of OFDM signal model is described for later chapters, the application environment of the model is explained. The modern signal analysis theory - cyclostationary analysis theory also is described, we analyze the cycle autocorrelation function (CAF) and the cyclic spectrum (CS) of common single-carrier communication signal and introduce its application in signal detection.Third, to explore the application of CAF for the OFDM signal recognition and parameter estimation, we analysis the periodicity of the autocorrelation function (AF) for the OFDM signal, the AF of different delay has different period, the cyclostationarity is introduced respectively by the over-sampling chip and the cyclic prefix, so it has multiple cyclic frequencies. And The CAF expressions of OFDM signals are theoretically derivated under different delay. Through analyzing the three-dimensional structure of the CAF, which shows that we can jointly estimate the chip time, symbol useful time and the symbol duration parameters of OFDM signals by searching the distance from the peaks in the corresponding profile. Finaly, we propose a numerical calculation method based on frequency domain accumulation for CAF of OFDM signals with low SNR. The high performance of the method to estimate the parameters of OFDM signals with low SNR is proved by simulations. However, the algorithm has large computation, it is still traditional autocorrelation algorithm for estimating the useful symbol time, and estimation of symbol duration is decided by the accuracy of the useful symbol time estimation, and the robust is poor under the conditions of multi-path and the short cyclic prefix OFDM signal. To solve these problems, we construct the cost function based on merging the multi-parameter, estimate the useful symbol time and the symbol duration simultaneously, through computer simulation, the recognition performance of the cost function algorithm has been greatly improved than the correlation algorithmsFourth, to explore the application of cyclic spectrum (CS) for the OFDM signal recognition and parameter estimation, the CS introduced by the cyclic prefix of OFDM signal is detailed. because of its numerical implementation requires very high spectral resolution, the calculation is very large, and the CS line introduced by the cyclic prefix is weak compared to the power line, the general fast numerical algorithm can not work. So the cyclostationary signature algorithm is proposed by selecting data symbol mapping of the sub-carrier set, and the rules of data set selection is proposed, the CS closed expressiones of the cyclostationary signature are derived. The algorithm greatly reduces the complexity of detection and identification , improves the speed of detection and identification, it can be used in cognitive radio system.Fifth, for the OFDM signals with short cyclic prefix or without cyclic prefix, minimum kurtosis algorithm, direct maximum likelihood algorithm, Gaussian maximum likelihood algorithm, matched filter algorithm, periodicity frequency kurtosis algorithm are proposed to estimate the parameters of the OFDM signal and identify the OFDM signal. And the approximate calculation method of direct maximum likelihood algorithm, Gaussian maximum likelihood algorithm and matched filter algorithm is given. The first four kinds of algorithms need to be synchronized, the synchronization steps can also be integrated into the algorithm. The characteristics of these algorithms are not sensitive to cyclic prefix and frequency selectivity channel. The performance of the algorithms are compared under different SNR. Finally, we test these algorithms with the real OFDM baseband signal collecting by our experimental platform. The results show that these algorithms can accurately estimate the parameter and identify OFDM signal, have great practical value. in cognitive radioSixth, to acquire the direction of OFDM signal in cognitive radio, to locate and track the target using non-cooperative emitters based OFDM signal, and other issues, It needs for DOA estimation of the OFDM signal. After analyzing the existing problem of DOA estimation for wideband OFDM signal by the cyclic MUSIC algorithm, Spectral correlation signal subspace fitting algorithm and average cyclic MUSIC algorithm, the MUSIC algorithm based on CS is proposed, combined with spatial smoothing algorithm to solve DOA estimation for wideband OFDM coherent signals, simulation show that the algorithm takes advantage of the selectivity of the cyclostationarity, it has good performance for wideband OFDM signals under the conditions of multi-path and low-SNR.