Tetrahedron-Constraint Least Square Localization Algorithm in Mixed LOS/NLOS Scenario Based on TDOA Measurements

The TDOA-based indoor localization is widely utilized due to high precision. However, the ranging signals may be contaminated by non-light-of-sight (NLOS) propagation, with the results that the accuracy and robustness of localization system can be seriously degraded. In this paper, a tetrahedron-constraint assisted localization algorithm is proposed to mitigate NLOS effects. The errors induced by NLOS in ranging are positive and larger than LOS measurement noises. Depending on the character, a range estimation based on Cayley-Menger Determinant (CMD) is presented to eliminate NLOS errors and obtain distances information from every 3 TDOA measurements. A group of position estimations are calculated by Least Square algorithm. Finally, a minimum weighted cost function is established to evaluate the candidate estimations in aspect of trace smoothing and estimation residual. A practical experiment verifies its better performance of accuracy and robustness.