H-BLADE: A Bayesian Probabilistic GNSS/LTE-OTDOA Hybrid Localization Algorithm for Harsh Environments

For global navigation satellite systems (GNSS), performance in harsh environments, such as urban canyons and indoors, can be improved by leveraging cellular localization systems. In this paper, we propose an algorithm, known as the Hybrid Bell Labs Localization Algorithm with channel bias Distribution Estimation (H-BLADE), which fuses measurements from GNSS and Long Term Evolution (LTE) - observed time difference of arrival (OTDOA) systems. The algorithm accounts for channel bias errors with a single heuristic probability distribution for each network. To provide additional robustness, the algorithm excludes the outlying measurements based on a metric called Mahalanobis distance. Using actual over-the-air measurements, we show that H-BLADE outperforms previous hybrid localization algorithms for GNSS and LTE-OTDOA that are based on either well-known nonlinear least square (NLS) or probabilistic techniques.