Study On Passive Localization Based On Time-varying Time Delay And Doppler Using Multiple Moving Sensors

Passive localization studies the geolocation of non-cooperative emitters by exploiting the properties of intercepted signals.Among the localization approaches,time delay and Doppler based localization determines the emitter's location by exploiting the difference of signal transmission delay and Doppler effect between intercepted signals.Time delay and Doppler based localization finds wide applications when relative movement between emitter and receivers exists.Plenty of works concentrated on the extraction of time delay and Doppler,the position determination from time delay and Doppler measurements,and the direct position determinations approaches.However,in scenarios with high-speed relative movement or long observation time span,time delay and Doppler based localization faces with new challenges as these parameters may change with time significantly.In response to the particularity of such scenarios,this work studies on the time-varying time delay and Doppler based localization,focusing on the correction of time delay variation,the compensation of doppler velocity variation,the extraction of time varying parameter in localization of stationary emitters,and the localization approaches using uncalibrated receivers.The correction of time delay variation in joint estimation of time delay and Doppler of unknown wideband signals is studied in chapter 2.Along with the movement of the sensors,signal transmission delay varies with time.The necessity of compensation of time delay variation is more significant for wideband signals.The existing approaches for wideband signals generally require multiple time-scaling of observed signals,resulting in almost prohibiting computational complexity.Alternatively,a joint estimation approach for wideband signals is proposed in this work,which could be illustrated as the coherent summation of the conjugate multiplication of short time intervals.With well compensation of time delay variation,the estimation accuracy of this approach could achieve the Cramer-Rao lower bound.To lower the computational complexity,a fast estimation algorithm based on Keystone transform is also proposed.After correcting the time delay variation through performing the Keystone transform,this fast algorithm could achieve the coherent summation of observed signals through the 2-D Fourier transform,lowering the computational complexity effectively.The compensation of Doppler velocity variation is studied in chapter 3.For scenarios with high-speed moving sensors or long observation time span,the Doppler velocity variation may be significant.In such cases,introducing the Doppler rate into the signal model might enhance system performance effectively.In the existing approach for joint time delay,Doppler and Doppler rate estimation,the effect of time delay variation on signal baseband envelope is neglected,might resulting in performance loss.For this defeat,an approach for joint time delay,Doppler and Doppler rate estimation based on segmentation of observed signals is proposed.To lower the computational burden,a fast estimation algorithm based on the sequence reverse transform and the second order Keystone transform is also presented.By performing such transforms mentioned above on the conjugate multiplication of intercepted signals in frequency domain,time delay,Doppler and Doppler rate is estimated separately,lowering the computational burden effectively.The time-varying parameters estimation in passive localization of stationary emitters is studied in chapter 4.Exploiting the fact that the emitter is stationary and located on the surface of the earth,equality constraint between time delay,Doppler velocity and Doppler rate exists in passive geolocation of stationary emitters,which might be exploited to enhance system performance.The expression of this constraint in airborne and spaceborne passive geolocation scenarios is derived.Based on this expression,the joint estimation algorithm considering constraint between parameters is proposed and the CRLB analysis is presented.Due to the information introduced by the constraint between parameters,higher estimation accuracy can be achieved with lower computational complexity.The direct position determination approaches using multiple uncalibrated receivers is studied in chapter 5,in the sense that each receiver applies an unknown complex scaling to the received signals and the unknown noise powers are possibly unequal.Taking the uncertainty of these parameters into consideration,the signal model using uncalibrated receivers is presented.The maximum likelihood estimate is derived and the CRLB analysis is presented.Different with the existing algorithms,normalization of the received signals is introduced in the proposed approach to whiten the receivers into unit power.Simulation results show that higher estimation accuracy could be achieved through the proposed algorithm in cases with unequal receiver noise powers.