The Trilateral Iterative Method Integrating LOS/NLOS Environment Recognition and Classification Ranging Error Suppression

Existing ultrawide-band (UWB)-based localization methods are likely to overlook environmental diversity and employ a single data processing approach to suppress ranging errors, which struggle to meet practical requirements. In order to address the issue, this article introduces the Trilateral Iterative method integrating line-of-sight/nonline-of-sight (LOS/NLOS) environment recognition and classification ranging error suppression. First, by calculating the height of the localization tetrahedron using Cayley-Menger determinant (CMD), the LOS environment threshold is established for environment recognition, reducing the impact of base station distribution changes on environment recognition and enhancing accuracy. Second, the Polynomial Fitting and Gaussian mixture model (GMM)–expectation maximization (EM)–interacting multiple model (IMM) are applied to suppress ranging errors in LOS and NLOS environments, respectively. The latter integrates GMM, EM, and IMM. Finally, the trilateral iterative method is proposed, which iteratively refines localization accuracy by continuously updating and reducing the error triangle based on the positional relationship between its vertices and center of mass. Experiments were conducted in the $9\times 10$ m area. The results demonstrate that the proposed method achieves an environment recognition accuracy of 92.34%, and the mean errors (MEs) of ranging are reduced by 96.10% and 87.90% in LOS and NLOS environments, respectively; furthermore, the MEs of localization are 25.93 and 75.77 mm in static LOS and NLOS environments, respectively, and 71.16 mm in dynamic environment. By accurately identifying LOS and NLOS environments, effectively suppressing ranging errors, and reducing the error triangle, the proposed method ensures the accuracy and stability of localization.