Diversity Based On Joint Time-frequency Processing In Spread Spectrum Communications

With the development of communication technology,the propagation path of radio waves is becoming more and more complicated.In most scenarios,the channel exhibits characteristics of double dispersion.The fast time-varying characteristics of the channel make it more limited to research the channel characteristics in the single time domain or frequency domain and to propose algorithms to suppress the channel fading.The time-frequency analysis is good at processing time-varying signals,and it has innate advantages in researching time-varying channels.Some scholars have proposed equalization algorithms in joint time-frequency domain to suppress the channel fading.In this paper,time-varying characteristics of fast fading channels are researched by common time-frequency representations.Time-varying spectrums and power spectrums of channels are also presented.Compared to the impulse response in time domain and the spectrum and power spectrum in frequency domain,this method reflects frequency components and amplitudes corresponding to each moment of the channel more clearly.The fast fading channel caused by Doppler spread makes the performance of diversity decrease.This paper presents the concept of joint multipath-Doppler diversity.Both multipath delay and Doppler spread can generates spread spectrum gains by diversity.After disintegrating the channel by short time Fourier transform,the processed signal is linearly combined by the time-shifted and frequency-shifted signals of the received signal.However,above operation depends on the autocorrelation of the spreading code.The improvement is less in the case of small Doppler shifts and the number of Doppler diversity is also affected by the spread spectrum gain.To solve these problems,this paper researches the diversity receiving method in joint time-frequency domain based on maximum ratio combining with best performance.This paper analyzes the sequence of the received signal and the estimated channel in joint time-frequency domain by short time Fourier transform.According to the convolution theorem in frequency domain,these two time-frequency distribution matrixes is linearly convoluted in columns.Some Doppler components in high frequency domain are also removed.The superiority in suppressing Doppler shift of this algorithm is verified from the derivation of theoretical formulas,the analysis of time-frequency diagrams and the comparison of error rate.In order to optimize the performance of the algorithm,this paper gives the type of window function and the window length with excellent performance based on the normalized equivalent noise bandwidth and the measure of aggregation.At the same time,considering the computational complexity of the time-frequency analysis,this paper reduces the overlap rate of window function while maintaining the performance of the algorithm.After deriving the relationship between the loss and the overlap rate,the minimum overlap rate is limited to about 75% based on the standard of bit error rate.