Hybrid ToA and Fingerprinting Localization Enabled by Channel Knowledge Map
Yuxiu Zhou, Chaoyue Zhang, Xiaoli Xu · 2024
Time-of-arrival (ToA) and fingerprinting are two classical approaches for indoor localization. With line-of-sight (LoS) links, the ToA information can be used to correctly estimate the range between the transmitter and the target. However, the accuracy of ToA localization degrades drastically if some LoS links are blocked. On the other hand, fingerprinting localization relies on the variation of signal features at different locations, which makes it suitable for rich scattering environment. However, due to the environment dynamics and measurement noise, fingerprinting can hardly achieve high localization accuracy, especially if only sparse and simple fingerprints are available, e.g., the received signal strength (RSS). To benefit from both approaches, the paper proposes a hybrid ToA and fingerprinting localization enabled by the channel knowledge map (CKM). Specifically, we construct a channel gain map (CGM) based on sampled RSS measurements, which is used as the fingerprints for coarse localization. Besides, the link state map (LSM) is also constructed based on the prior physical environment information and the CGM, which stores the LoS link status at various locations within the area of interest. Given the coarse target location, those transmitters that have LoS links with the target are identified based on the LSM, and then ToA information from those transmitters is used to refine the localization. Simulation results show that the proposed hybrid localization scheme can significantly outperform that benchmark schemes.