Study On Adaptive Equalizer In Shortwave Communication

With the characteristics of far transmission distance, simple structure and the communication medium not being destroyed, shortwave communication is extensively applied in military and emergency communication. In shortwave communication, the quality and speed of communication are seriously affected by the intersymbol interference caused by multi-path transmission deriving from the characteristic of the channel. At present, the adaptive equalization is one of the main technical means to overcome multi-path effects and reduce inter-symbol interference, which essentially filter the received demodulated signal by adaptive filter. On the basis of analyzing characteristics of shortwave channel, this paper studies the advantages and problems of several equalizer algorithms and structures, and then conducts the function simulation of the LMS algorithm. The main work of this paper are as follow:Firstly, this paper introduces the application in communication system and the history of Adaptive Algorithm briefly. The characteristics of shortwave and the reasons of intersymbol interference are studied. Based on the Nyquist theorem, the time domain equalization is described in details. This paper establishes the Watterson mathematical model of shortwave, and determines the channel's parameters according to ITU-R.Secondly, this paper conducts extensive theoretical researches on several kinds of equalizer structures. Based on Watterson shortwave channel, this paper applies the least mean square algorithm and recursive least square algorithm, and conducts both theoretical research and computer simulation to study the variables which influence the convergence rate, the tracking ability and the steady-state error of the algorithm. The simulation results proved that, in the same channel environment, compared with LMS algorithm, the RLS algorithm is one order of magnitude faster in the aspect of the convergence rate, and the effect of RLS algorithm is better in the aspect steady-state error. But, RLS algorithm's computational complexity is higher than LMS algorithm's. The LMS algorithm is more simple and its engineering realization is easier. In order to reduce the instability of LMS algorithm's steady-state error, the Normalized LMS is studied from the aspect of step factor. It is proved by the simulation that, compared with LMS algorithm, NLMS algorithm can keep steady-state error better when the signal is large. To accelerate the convergence rate, this paper studies Decorrelation LMS algorithm and Transform domain LMS algorithm through theoretical derivation and simulation. It is shown by simulation that the improved LMS algorithm'convergence rate is much faster than that of the traditional LMS algorithm.Thirdly, this paper conducts the function simulation of LMS-type adaptive equalizer based on FPGA. To solve the problems existing in the floating point number simulation, i.e., complicated arithmetic and time consuming, by the premise of guaranteeing the precision of operation with Finite Word Length Fixed-Point, fix point simulation is adopted to realize the equalizer, whose advantages is fast operation and easy realization. In the design of FPGA, modular design and serial-parallel structure are used to implement the LMS-type adaptive equalizer. The results of simulation have proved that the improvement scheme effectively reduced inter-symbol interference and with small calculating amount and low occupancy ratio of resources. The curve of mean square error of the scheme is consistent with the theoretical analysis, which suggest that the results of this improvement scheme have good practical value.