Research On Localization Against Marine Environment

Ocean space is essential to the national economy and military.Underwater wireless sensor networks(UWSNs)as an important means of marine information acquisition becomes a research hotspot in recent years.Since the satellite positioning system is infeasible for underwater scenarios due to different transmission medium,Using UWSNs for obtaining the location information of an underwater target is becoming ubiquitous in oceanic applications,such as ocean environment monitoring,disaster prevention and underwater navigation.In the underwater environment,UWSNs have the characteristics of mobility,sparsity,energy-constrained and harsh underwater channels.Traditional localization methods which are used in terrestrial wireless sensor networks cannot be directly applied underwater,this makes underwater localization facing significant challenges.Based on measurements including time difference of arrival(TDOA)or frequency difference of arrival(FDOA),this paper studies the localization principle,localization methods and localization performance analysis under different localization scenarios.The main contents and results are as follows:(1)Cyclic feature-based Doppler factor estimation method for underwater acoustic channels and the Doppler-assisted localization method with synchronous clocks.First,we analyze the cyclic features of cyclic prefix(CP)orthogonal frequency-division multiplexing(OFDM)signals over doubly selective underwater channels in order to demonstrate the relationship between the cyclic features and the Doppler scale.Based on this relationship,we propose a Doppler scale estimation approach exploiting the cyclostationarity of CP/zero padding(ZP)-OFDM signals.Our proposal is simple to implement and needs no training data or specially designed packet structure,which improves the channel spectrum efficiency.Then,based on the measured Doppler scale constructed FDOA measurements and the TDOA measurements,we propose a two-step weighted least square(WLS)estimator for joint estimation of the target position,target velocity and sound propagation speed with unknown propagation speed and sensor parameter errors.The theoretical effectiveness of the new localization method is analyzed and proved.Simulation studies are performed to verify the effectiveness of the proposed localization method.(2)The Doppler-assisted localization method with asynchronous clocks.We first point out the problem of the extraction of differential Doppler scale(DDS)and TDOA measurements from the ranging signals with asynchronous clocks.From the received signal model,it is found that the DDS/TDOA measurements are tightly coupled with the clock skew.Aiming at this phenomenon,a new DDS/TDOA model with asynchronous clock is established.Based on this model,a two-step WLS method for joint estimation of the target position and clock skew is proposed.The method is simple and effective.The simulation results show that the parameter estimation performance of the proposed localization method reaches the Cramér-Rao Lower Bound(CRLB)under the condition of small noise.(3)The localization method using ray tracing under the condition of sound speed uncertainty.First,the stratification effect is modeled by using the ray tracing theory and the joint synchronization and localization problem is derived.The corresponding maximum likelihood(ML)estimator and the CRLB are derived.The ML estimator is shown to be highly nonlinear and nonconvex,therefore,we employ the Gauss-Newton algorithm to solve the original nonconvex ML problem in an iterative manner.In order to launch the Gauss-Newton method,we need a suitable initial position that is sufficiently close to the optimal solution.We assume the target node has been synchronized and the sound speed in underwater environment is a constant value,the initial position is calculated using LS estimator.The simulation results show that the performance of the proposed method is better than the existing methods and close to the CRLB.In terms of computational complexity,the proposed method also has great advantages.(4)Localization methods using equivalent sound speed model under the condition of sound speed uncertainty.Different from the previous assumption that the sound propagation path is a straight line,these methods use the equivalent sound speed to characterize the uncertainty of sound speed in the underwater environment.The equivalent sound speed model greatly simplifies the ranging/time measurement equation and lays a foundation for the use of lightweight localization methods(e.g.LS),but it introduces a new unknown constant,which needs to be jointly estimated in the localization problem.Firstly,a WLS-based joint estimation method of target position and sound speed is proposed.This method uses TDOA measurements obtained from the sea surface reflection multipath and direct path.The statistical effectiveness of the proposed method is verified by theoretical analysis and simulation experiments.Secondly,aiming at the defects of sparse reference nodes and insufficient ranging measurement in the underwater environment,an underwater target cooperative localization method based on belief propagation theory is proposed.This method is a distributed self-localization method,which has good scalability.The simulation results show that the proposed method has fast convergence,wide coverage and high localization accuracy.Thirdly,aiming at the low-power tracking problem of Autonomous Underwater Vehicle(AUV)in UWSNs,a tracking protocol considering the energy consumption of the whole network(including reference nodes and the AUV)is designed and the packet delivery success rate(PDSR)is analyzed.Since the low-power tracking protocol leads to the intermittent arrival of observations,this paper proposes two AUV tracking methods based on Intermittent Extended Kalman Filter(IEKF)and Intermittent Unscented Kalman Filter(IUKF),respectively.The simulation results verify the accuracy of the protocol theory analysis and the effectiveness of the proposed tracking method.The influence mechanism of the protocol parameter design on the tracking performance is also demonstrated.