A robust EM-SOCR localization method under probabilistic NLOS environments considering anchor node position uncertainties

In the field of integrated sensing and communication (ISAC), source node positioning is a key issue to achieve intelligent sensing and efficient communication. A new source node localization scene that both considers the probabilistic non -line -of -sight (NLOS) environments and anchor node position uncertainties is introduced. Localization is realized based on the measured distances, whose noise variances are supposed to be linearly dependent on path lengths. Firstly, an expectation maximization and second -order cone relaxation (EM-SOCR) procedure is proposed to eliminate the NLOS influences, and then the source node position is estimated. Practical experiments demonstrate the effectiveness of the proposed algorithm in significantly improving positioning accuracy, achieving a 33.8% performance enhancement in the test scenario.