Study On Sparsity-Aware Multi-Source Localization Using TDOA

Target positioning technology has been greatly developed and widely used in recent decades.TDOA-based positioning technology has attracted strong attention due to its high precision.The current research direction of TDOA positioning is mainly for a single target source,while the research on multi-source TDOA target positioning is relatively rare.Because the location of the target source in the space is unique,the space is sparse,and the sparsity-aware framework is used to precisely locate multiple target sources.The main work of this thesis is as follows:1.For sparsity-aware technology.The sparsity-aware techniques and their properties under different norms are described,and the restricted isometry property that the reconstructed signal needs to satisfy are discussed.And for OMP algorithm and LASSO algorithm,the effects of sparsity and measurement dimension on reconstruction performance are discussed through simulation experiments.2.For single-source TDOA positioning.Since the TDOA positioning model is a nonlinear and complex equation,three classical single-source TDOA positioning algorithms are introduced,namely Taylor series method,Chan method and cwls method,and simulation experiments are carried out for the three algorithms.3.In the single-target source TDOA positioning simulation experiment,various algorithms have shown relatively good positioning performance.For multi-source TDOA positioning,TDOA needs to allocate corresponding sources,so it is more complicated than single-source TDOA positioning.By sparsifying the space and using a novel TDOA fingerprint grid design model,the multi-source TDOA localization problem is transformed into a norm minimization problem under the sparsity-aware framework.4.In order to improve the ability to locate the number of target sources,inspired by the increase in the dimension of measured values in sparsity-aware,nonlinear equations can be used to process the sparse equations,so as to accurately locate multiple targets at the same time.In some special cases,the underdetermined equation can be transformed into an overdetermined equation,which can be solved directly by the classical least squares method.5.For the case where multiple target sources are not on the grid,it can be regarded as a grid mismatch problem.The perturbation matrix is added to the sparse equation,and then it is discussed,and two solutions are given.Finally,simulation experiments for multi-source TDOA localization are carried out,and a large number of simulation results prove that the proposed model has good performance.