Study On UWB Positioning Method And Configuration Optimization

With the rapid development of navigation and positioning technology,the importance for high-precision location information is increasing.Satellite positioning technology has the characteristics of providing three-dimensional positioning all-weather,which can meet the needs of outdoor positioning,but in indoor environment,high-precision positioning cannot be achieved due to the obstruction of buildings.UWB positioning technology has the advantages of strong penetration,low power consumption,and anti-interference,and can be applied to the positioning,tracking and navigation of indoor objects and people.Therefore,it is of great significance to study the relevant theories and methods of UWB positioning,establish a more reasonable and effective indoor positioning system,and provide accurate indoor location services.Hence,this paper studies the UWB positioning method and configuration optimization.The main research contents are as follows:(1)UWB technology and characteristics,synchronous and asynchronous positioning application systems are introduced,angle and distance measurement methods are compared,positioning models and parameter estimation methods are discussed,and the performance of several iterative methods are compared and analyzed.As the measurement process of UWB positioning system is affected by a variety of different mediums,the UWB signal's ranging error characteristics are statistically analyzed.At the same time,the modeling compensation of ranging error is realized based on the changing law of ranging error,and the accuracy of ranging positioning is improved.In addition,the low-frequency noise belonging to the multipath error and the high-frequency noise of the random error are separated based on the empirical mode decomposition method,and the influence of the multipath error on the UWB positioning is analyzed.(2)For the positioning model solution based on the Gauss-Newton iteration method,the linearization process of the distance equation ignores the influence of higher-order terms,which causes model bias.The influence factors and laws of the second-order model items of Taylor expansion on the parameter estimation bias are analyzed.Then,the hypothesis test of the parameter bias is carried out to quantify the linearization model and judge whether the model accuracy is sufficient for UWB positioning and whether the parameter bias can be ignored.The results show that the parameter bias is determined by the accuracy of priori observation,the design matrix of the positioning geometry and the nonlinearity.For relatively high measurement accuracy,the linearization is effective,and the parameter bias can be ignored.At the same time,when the nonlinearity and the design matrix are orthogonal in the vector space,that is,the anchors are uniform and reasonable,and the parameter estimation tends to be unbiased..(3)Due to the limitations of indoor environment,it is not easy to obtain good initial values based on the Gauss-Newton iteration method for UWB positioning.Especially when the positioning system is ill-conditioned,it can not converge to the global optimal solution or even diverge.By introducing the closed-Newton iteration method that considers higher-order terms,and comparing it with the Gauss-Newton method,it is shown that the closed-Newton method has more stable and effective convergence and better positioning performance under the condition with increasing a certain amount of calculation.Particularly when the positioning system has ill-conditioned problems or relatively low observation accuracy,no matter whether the initial value is reasonable or not,it can better converge to the global optimal solution.(4)Nonlinear filtering algorithms have been widely used in navigation and positioning.However,the outliers in the measurement process make the innovation vector increase abnormally,which will affect the filtering performance and reliability.Therefore,two robust estimation methods of nonlinear filtering are proposed to be applied to UWB dynamic positioning.The global test will be carried to to judge the validity of the model.REKF uses the IGGIII scheme to determine the single outlier based on local test of the normalized residual,and then realize residual optimization and robust estimation.RUKF reduces the anomalous influence of outliers by adjusting the measurement noise covariance.The results show that when the outliers are treated as systematic errors and large variance random errors,REKF and RUKF can effectively resist and weaken the influence of the outliers and improve the positioning accuracy.When the outlier comes from the thick-tailed distribution,at this time,the least square estimation is not the maximum likelihood estimation,the detection based on the chi-square test is not reasonable,and the robust estimation is ineffective.Then,the maximum likelihood estimation based on the L1 norm estimation should be considered.(5)Positioning parameter estimation is related to observation quality,positioning method and spatial geometric configuration.GDOP is widely used to evaluate the performance of navigation and positioning systems,and the condition with GDOP minimization is also suitable for UWB positioning configuration optimization.Based on whether clock errors are considered,two types of GDOP minimization conditions are introduced,and the cone configuration with lowest GDOP is introduced and analyzed.The results show that for two-dimensional positioning,a regular polygon is the best configuration.It not only achieves the minimization of the second type of GDOP,but also compensates for various systematical errors.In three-dimensional positioning,signal transmission limits the positioning configuration.Meanwhile,the two-way ranging technology is widely used in the measurement for UWB positioning,it makes more reasonable to optimize the configuration based on the first type of GDOP minimization.In addition,by adding some constraint information,the configuration optimization of global area can be more effectively realized.