Identification And Compensation Processing Of Nonlinear Distortion In Wide-Band Receiver Front-End And Its Application

SDR techonology meets the increasing demand of wireless communications in complex electromagnetic environments.In order to give full play to the advantages of SDR,the flexible configuration of software back-end,the receiving front-end must be wide and large dynamic range system to realize total probability of receiving the full band signals.However,nonlinear distortions generated by the cascaded nonlinear devices(such as amplifiers,filters,mixers and ADC)in the wideband receiver badly affect its spurious-free dynamic range(SFDR),which further affects the reception performance of some weak in-band signals.The key to the problem is to improve the SFDR of the wideband receiver.At present,the main mean to improve the SFDR of receiver is to reduce or even eliminate the nonlinear distortions introduced by those nonlinear devices in the wideband receiver through digital post-compensation.However,the frequency-domain blind identification compensation methods are not only applicable to the narrow-band receiver,but also have high computational complexity and are not real time compensation methods.Thereby,two time-domain digital post-compensation algorithms are proposed to enhance the linearity of the wideband receiver.Meanwhile the method,single-channel identification and multi-channel synchronous compensation,is proposed to improve the linearity of array receiver.In addition,a new channel encryption technique is proposed and verified by the simulation and experimentation results based on active nonlinear transformation.The technology utilizes the nonlinear distortion that is actively added to create obstacles for the non-partners to receive and judge the encrypted signal,enhancing its anti-interception capability.The main contributions of this thesis can be summarized into the following four aspects:1.A time-domain nonlinear blind identification and compensation algorithm is obtained by quantificationally mathematical calculation and theoretic analyses of the nonlinear distortion mechanism of single-channel broadband RF receiving front-end.We adopt the SRA-TFC method to extract small signals(mainly consisting of nonlinear distortion components),and large signals(high-powered frequency components)fromoutput signal of receiver front-end in time domain,respectively.Additionally,a blind identification criterion of the nonlinear behavior model is designed as follows: minimizing the residual sum of squares between the small signals and nonlinear model of large signals.Moreover,the improved weighted iterative algorithm is applied to extract and update the parameters of the nonlinear behavior model.Finally,the updated and extracted model kernel coefficients can achieve to compensate the nonlinear distortion of the receiver in real time.The simulation and real-world experimental results both show that the SFDR of the receiver achieves 15-20 dB improvement,leading to enhance the ability to receive and detect weak signals in concomitance with strong jammers.2.This paper proposes a novel compensation technique aiming at the nonlinearity mitigation of array receiver and the SINR improvement of weak signals covered by the nonlinearity,which fundamentally leads to a high and stable estimation accuracy of the array signal.Moreover,we offer the feasibility analysis of the suggested method which adaptively extracts the nonlinear model kernel coefficients of any one of the channels to mitigate the nonlinearity of all channels synchronously.The shortwave array signal used for testing is from an actual uniform circular array,the obtained results indicate that the SFDR of the entire array receiver in the full band increases by 5–15 dB,improving the2-D DOA estimation performance of weak target signals.3.Based on LS-SVM,the proposed nonlinear compensation technique is applied to wideband digital RF front-end.This technology utilizes the excellent fitting ability of LS-SVM to identify and model the nonlinear inverse system,which enables the input and output of the entire RF receiving front-end to tend to be linear mapping.Firstly,the technology needs to construct a suitable training sample set.Moreover,the LS-SVM inverse model of the RF receiving front-end is obtained by LS-SVM regression algorithm.Then the hyper-parameters of the inverse model are optimized and the optimal parameters of the inverse model are solved.Finally,the output signal of RF receiving front-end is used as the test sample to compensate its nonlinear distortion components.The simulation and real-world experimental results both display that the SFDR of the receiver achieves about 20 dB improvement,increasing the ability to receive and detect weak signals.4.A physical layer channel encryption technique based on active memory nonlinear transformation is proposed,which is different from the traditionalsecure techniques in communications such as source encryption techniques.This technique actively generates strong nonlinear distortions by nonlinear models before the encryption information is transmitted to make it difficult to intercept,recover and decipher the original encryption information for non-participants.But the cooperative receiver performs inverse transformation on the known nonlinear model to accomplish the elimination of active nonlinear distortion components and then uses the decryption key to restore the original signal exactly.First,the principle and steps of the proposed technology are described in detail.Second,three nonlinear models are designed,and their invertiblity are demonstrated theoretically under the condition of strong nonlinearity.And on this basis,an optimization mechanism is designed to enhance the anti-interception ability of the proposed technology according to the loop transformations of three nonlinear models.Finally,the validity and security of the suggested technology are verified by the simulation and experimentation results.