Virtual Antenna Array and Multipath AOA-Delay Fingerprints Based Location for Moving Targets

• Main Authors: Zhigang Chen, Lei Wang, Lu Chen
• Format: Article
• Language: English
• Published: IEEE 2020-01-01
• Series: IEEE Access
• Subjects: Fingerprinting localization; multipath AOA-delay fingerprints; channel frequency responses (CFRs); two-dimensional inverse Fourier transform (2D-IFT); wireless positioning
• Online Access: https://ieeexplore.ieee.org/document/9217431/

Multipath effects are the major challenge to the accurate indoor wireless positioning technology and the most viable solutions for overcoming such multipath effects are received signal strength (RSS) or channel status information (CSI) based fingerprinting location methods. However, these fingerprinting methods are usually subject to the RSS/CSI fingerprints' high variability over space. In this paper, a novel virtual antenna array and multipath angle of arrival (AOA)-delay fingerprints based location method has been proposed. The proposed method employs 2-dimensional inverse Fourier transform on the CFRs (Channel Frequency Responses) at equi-spaced positions on a quasi-straight line moving trajectory, i.e., the CFRs of a virtual uniform linear array (VULA), to extract the equivalent multipaths' AOAs and delays as the radio fingerprints, which vary much more slowly than the RSS/CSI fingerprints in the space domain. By exploiting the geometric properties of the multipaths' AOAs and delays, a complete fingerprint database containing fingerprints of moving trajectories in different directions with each RP (reference position) as their midpoints can be constructed based on just two fingerprints of two perpendicular line trajectories with the same RP as the midpoints. Further, a two-step online fingerprinting location method is developed, the multipath delay estimates are used to narrow the scope of positioning in the first step, a modified WKNN (Weighed K-Nearest Neighbors) fingerprinting method, with the reciprocals of the fingerprints' Euclidean distances adopted as the weighting factors, is then employed on RPs in such a narrowed positioning scope to enhance the positioning accuracy. Experimental results verify that the proposed fingerprinting method outperforms the corresponding traditional RSS/CSI fingerprints based approach.