Multiple-epoch joint localization and synchronization in a 5G system

Currently, the usage of GNSS signals for positioning in challenging propagation conditions like urban and indoor scenarios is still an open problem. The 5G network-based localization technique can help relieve the problem of lacking visible satellites in urban environments. Even it can work as an independent positioning means in the GNSS denied environments. However, in real applications, due to the limited cost, the 5G BSs cannot be equipped with expensive atomic clocks to keep time. Thus, the 5G inter-BS synchronization error cannot be ignored simply, which becomes a main limiting factor for 5G-based positioning techniques. To address the impact of 5G inter-BS synchronization error on positioning accuracy, this work proposes a multiple-epoch jointly localization and synchronization algorithm. We model the synchronization error of each BS as an unknown quantity to solve. This makes it possible to estimate the synchronization error of each BS precisely and finally get rid of the effect. Then we employ the correlation feature over multiple epochs to reduce the overall number of quantities to estimate. The Taylor series least square method is extended to fuse time of arrival (TOA) - angle of departure (AOD) measurements for positioning scheme. Further, the lower bound of positioning error of the proposed method, namely Cramer Rao Lower Bound (CRLB), is derived mathematically. Results show that the proposed method significantly reduce the positioning error compared with three baseline methods. It achieves joint localization and synchronization to improve the positioning accuracy in an imperfectly-synchronous 5G network.