Research On Several Key Techniques For Wideband Receivers

Compared with traditional narrowband receivers,wideband receivers have broader receiving band,which endows the receivers with better performance in multi-mode adaptation,improving the robustness of communication link,and wideband spectrum monitoring.Owing to these advantages,the widening of receiving band has become an important trend for communication and electronic warfare receivers.The key technics for wideband receivers include wideband sampling,wideband interference resisting,dynamic subsignal extraction,and wideband spectrum detection,which are just the subjects of this thesis.The aim of the thesis is to address the deficiency of current research.In detail,the thesis has done the following work:1.We have studied the hybrid filter bank(HFB)based wideband sampling.We proposed a modified method for HFB design,which can significantly improve the resolution of HFB compared with traditional methods,while the cost is only a slight loss in sampling bandwidth.Due to the realization error of analog filters,the HFB may not achieve its theoretical resolution.To solve this issue,we proposed a method to test HFB's performance and estimate analog filters' frequency responses.This method can intuitively reflect whether the HFB reaches the required resolution or not;if not,this method can accurately obtain analog filters' frequency responses,thus providing the basis for the calibration of HFB.We also proposed a kind of HFB that can adjust its analog-to-digital conversion band.This HFB can help the receiver reduce its resource consumption to some extent.2.We have studied the HFB based anti-interference technology for wideband receivers.We proposed a novel wideband receiver architecture,which can achieve good performance both in interference resisting and signal adaptation.The proposed architecture utilizes a row of analog filters with steep transition band to cover the receiver's whole working band.When the receiver is under operation,only the analog filters whose passbands have useful signals are turned on.The analog filters at which interfering signals locate are closed.In this way,out-of-band interferences can be avoided.On the other hand,the active analog filters serve as the HFB's analysis filter bank and a matching bank of synthesis filters are configured,to reconstruct the spectrum of useful signal.In this way,the proposed architecture is endowed with the ability of receiving signals with different bandwidths.3.We have studied the fast filter bank(FFB)based digital channelization technology.We proposed a criterion for FFB's channelization performance,which can quantitize FFB's ability of extracting dynamically changing subsignals without mutual interference.We modeled the optimization problem of minimizing FFB's computation complexity and group delay,under the restriction of a certain channelization performance.To solve this optimization problem,we derived the relationships between the target functions and the two designing parameters,i.e.,the FFB's level number and the prototype filter's transition bandwidth.Then we proposed a two-dimensional search algorithm to obtain the optimal parameters.Experiment results indicate that,in most scenarios,there exists an interval for parameters that make computation complexity and group delay approach their minimum values simultaneously.4.We have studied the wideband spectrum detection technology.We proposed a new wideband spectrum detection method which is based on the statistical analysis of spectrum values.The proposed method needs not to know the noise power for determining threshold,thus it avoids the traditional methods' demerit of being seriously affected by noise uncertainty.Simulation results indicate that the proposed method can achieve good detection performance under low SNR region,as well as obtain constant false alarm probability.The ideology of our method is that: Adopting time-interleaved ADC to realize the Nyquist sampling of the wideband,and calculating the spectrum of the high-speed sampled signal in a parallel way,through which multiple estimating versions of the useful signal spectrum can be obtained.The statistical characteristic of the correlation between these estimated spectrum versions is independent of noise power.Therefore,a detection statistic that is also independent of noise power can be defined—the difference of average correlation coefficient's cumulative distribution functions.