Fuzz C-Means Clustering Algorithm for Hybrid TOA and AOA Localization in NLOS Environments

• Published in: IEEE communications letters 2024-08, Vol.28 (8), p.1830-1834
• Main Authors: Zou, Yanbin, Zhang, Zekai
• Format: Article
• Language: English
• Subjects: angle-of-arrival (AOA); Clustering algorithms; fuzz C-means (FCM) clustering; Location awareness; Noise measurement; non-line-of-sight (NLOS); Position measurement; Sensor systems; Sensors; Simulation; Source localization; time-of-arrival (TOA)
• ISSN: 1089-7798; 1558-2558
• DOI: 10.1109/LCOMM.2024.3408297

In this letter, we investigate the localization problem of hybrid time-of-arrival (TOA) and angle-of-arrival (AOA) measurements in non-line-of-sight (NLOS) environments. It is well known that the NLOS errors are usually much larger than the measurement noise, and they could severely degrade the accuracy of localization systems. In hybrid TOA and AOA localization systems, only one sensor can determine the location of source, and multiple sensors can provide multiple probable location estimates for the source location. Using the fact that the estimates provided by line-of-sight (LOS) sensors are usually adjacent to the true source location and the estimates provided by NLOS sensors are usually far away from the true source location, we can utilize the Fuzz C-means (FCM) clustering algorithm to separate the estimates into two clusters, LOS and NLOS, respectively. Then the cluster with a larger membership grade sum is selected as the LOS cluster, and the center of the LOS cluster is selected as the estimate of source location. Finally, simulation results validate the performance of the proposed algorithm.