Research On Localization And Tracking Under The Influence Of Complex Factors

With the rapid progress in wireless communication,wireless sensor node,information pro-cessing and computation,wireless sensor networks(WSNs)and mobile communication networks(MCNs)have attracted many research interests from both the academic and industry communities,and are widely applied in various areas,e.g.,environmental monitoring,advanced manufacturing,E-health care,location based services,etc.Location information is essential for many applications in both WSNs and MCNs,which makes localization technique become one fundamental technique.As one of the most important applications of WSNs,target tracking plays a signification role in both civil and military area.Therefore,localization and tracking techniques have gained great attentions of global researchers.Algorithm design is one of the key issues in localization and tracking.A large amount of algo-rithms for either localization or tracking have been proposed in recent years.Many complex factors in practical environments,e.g.,non-line-of-sight(NLOS)propagation,multipath,etc.,bring new challenges for localization and tracking.Because of the impact of the complex factors on signals,the propagation properties of the signal will be changed,which makes the measurements much more complicated.The algorithms based on ideal assumptions will be neither working nor robust to complex factors.Although some algorithms have been proposed to tackle the localization and tracking under complex factors,they have some drawbacks in various aspects,e.g.,computational complexity,localization accuracy.etc.In this thesis,we make a comprehensive investigation of localization and tracking from different aspects by considering the impact of a few complex factors on the measurements.The main work and contributions are summarized as follows:1.A brief literature review on localization and tracking techniques and the main challenges of localization and tracking under the influence of complex factors are provided.2.Localization accuracy analysis and sensor coordination in target tracking are conducted based on a measurement model including both additive and multiplicative noises.Firstly,the Cramer-Rao Lower Bound(CRLB)based localization accuracy is derived,whose properties are analyzed accordingly.Then,the measurement model is verified by lab-scale experimen-t.Secondly,for target tracking in mobile sensor networks,under the same measurement model,a coordination strategy for mobiles sensors is proposed to maximize the tracking ac-curacy,including sensor selection and motion strategies.The performance of the proposed coordination strategy is verified by simulations.3.For the node localization problem in WSNs,a more general measurement model is con-sidered.We assume the measurement noise is unknown but bounded,and formulate the localization problem as a non-convex optimization problem to minimize the worst-case esti-mation error.Relaxation is applied to transform the non-convex problem into a convex one.Geometrical analysis providing additional insights is also provided.Furthermore,a distribut-ed algorithm is proposed,whose convergence is proved.The simulation results reveal the performance of both the centralized and distributed algorithms.4.The localization problem in probabilistic NLOS environment is investigated.A Maximum Likelihood Estimator(MLE)is proposed for localization,which incorporates the distance de-pendent LOS(Line-of-sight)/NLOS probabilities.Then,CRLB based on the general LOS/NLOS probability model and NLOS bias model is derived.Through numerical analysis,the impact of NLOS propagation on the localization accuracy is evaluated.The performance of the proposed algorithm is verified by real world experimental data.5.The data association problem in multiple target tracking is investigated.Due to the effect of target size,occlusions will occur during targets' movements.A three-step algorithm is pro-posed to identify the occlusions.To reduce the computational complexity of traditional Joint Probabilistic Data Association(JPDA)algorithm,a modified JPDA algorithm is proposed to execute data association by utilizing the information of occlusion conditions.Simula-tion results show that the proposed algorithm is with good tracking performance and low computational complexity.In the end,the thesis is concluded and some future research topics are discussed.