NLOS Position Estimate Algorithm Based On Optimization Method In NLOS Environment

With the development and progress of the information industry,in-door has become the space where most modern people live and work,and people's demand for indoor high-precision location services is increasing,so positioning terminals have an urgent need for high-precision and stable positioning algorithms.5G mobile communication technology can provide more accurate range and angle measurement accuracy in three major ap-plication scenarios.Therefore,5G puts forward new viewpoints and meth-ods for the application and development of wireless positioning technology in complex environment positioning.However,indoor positioning scenar-ios have complex building layouts,and wireless signals are prone to non-visual propagation due to signal refraction and reflection,which will cause large deviations in distance and angle observations between positioning anchor nodes and positioning terminals and seriously affect the perfor-mance of location estimation.In order to improve the ability of position estimation in harsh environments,this thesis conducts research on the TDOA(Time Difference Of Arrival)positioning algorithm in indoor scenes where line-of-sight and non-line-of-sight positioning environments coexist.The TDOA localization solution is a typical nonlinear nonconvex problem,so nonlinear least squares and convex optimization ideas can be adopted to optimizes the location estimation to further improve the locali-zation performance of the localization algorithm in complex signal propa-gation environments.The specific research process and contents of this thesis are as follows.First,for the localization environment where non-visual and visual anchor nodes co-exist in wireless localization,this thesis proposes an improvement strategy to the traditional Chan-Taylor combination position estimation method.First,based on the traditional residual weighting algo-rithm,the screening rule of anchor node combination is improved by the principle of minimum residual,and the higher-order term of the residual function is used in the weighting process instead of the original first-order term as the weight,so that the estimated results of the improved residual weighting algorithm are used as the initial values of the position estimation.Then,under the premise of obtaining the initial value of the location esti-mation,nonlinear least squares optimization is used to iteratively find the location points,and this initial value is first used to compensate the devia-tion of the distance difference observation in combination with the most rapid descent method,so as to construct a location target function closer to the true value,and then the target location is solved iteratively using the trust domain algorithm.Finally,the accurate results and rough residuals are weighted to derive the final positioning results.Then,a convex optimization-based position estimation algorithm is proposed for the harsh positioning environment where all base station an-chor nodes in wireless positioning have non-visual range errors.First,the quadratic programming idea and Chan's algorithm are combined to propose a TDOA localization algorithm based on quadratic programming.Then,in order to improve the position estimation performance of the method under non-line-of-sight conditions,the distance observation,the geometric rela-tionship between the anchor node and the target,and the most rapid descent method iteration are used as the constraints of quadratic programming,which improve the accuracy of TDOA localization solution to a certain ex-tent.Then,the quadratic programming localization model is equivalently transformed into a second-order cone programming model,and a new pen-alty function is added to the target equation to improve the problem of out-of-bounds real position search range in the original algorithm,which has a higher solution efficiency compared with quadratic programming and thus improves the stability of the localization solution.Finally,this thesis evaluates the performance of the improved posi-tioning algorithm through the simulation.Compared with the traditional positioning method,the positioning accuracy and stability are improved to a certain extent.