Continuous TOA Measurement-Based Single-Anchor Indoor Localization

• Published in: IEEE access Vol. 13; pp. 122568 - 122580
• Main Authors: Lei, Ting, Pan, Shengshan, Wang, Kaiyu, Yu, Yan
• Format: Journal Article
• Language: English
• Published: IEEE 2025
• Subjects: angle estimation; Antenna arrays; Blades; Estimation; line-of-sight recognition; Rotation measurement; support vector machine (SVM); Support vector machines; Time measurement; time-length distribution; Wireless communication; Wireless localization
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2019.2941596

On the premise of guaranteeing high-quality measurement data, active wireless localization (A-WL) locates the target by at least three anchor nodes (ANs) with known coordinates, while passive wireless localization (P-WL) requires more ANs. In addition, P-WL requires a cumbersome training process to maintain the robustness in application scenarios as much as possible. However, the optimal deployment of ANs is not an easy task, and the application scenario will not remain unchanged. Therefore, a lightweight localization system that satisfies the localization accuracy requirements and can be quickly deployed should be one of the preferred solutions for WL. In this paper, we propose a single-anchor localization system based on continuous time-of-arrival (TOA) measurement. Specifically, we fix the AN on the blade with uniform rotation, construct the time-length distribution map by using the cyclically varying TOA caused by blade rotation, and intercept it according to the rotation period. We extract the change angle and distance information from the intercepted data in real time to achieve single-anchor localization. Meanwhile, we also construct a support vector machine classifier to identify the line-of-sight state of each interception period to assist the angle estimation in harsh environments and further improve the adaptability of the system. Experimental results show that the line-of-sight recognition accuracy of higher than 92% under continuous movement, and the average localization error of less than 0.6m. The average localization error is less than 0.3m in the mixed motion state.