Clustering Data Association Using Data Relevance in Spatial Domain for Doppler-Only RSN Localization.
Qianli Wang,Zhuming Chen,Zisheng Wang,Zhiqin Zhao,Zhengyang Wu
DOI: https://doi.org/10.1109/taes.2018.2836758
IF: 3.491
2018-01-01
IEEE Transactions on Aerospace and Electronic Systems
Abstract:Localization of multitarget by means of a network of dispersed radar sensors using Doppler-only measurement faces the problem of data association, i.e., associating the measured Doppler-shifts with their originated targets. Conventionally, this association is realized based on correlations between successive estimations of targets' state. However, up to now, no data association method is available for localization using only one snapshot measurement. This paper tries to utilize the relevance of measurements in spatial domain instead of in time domain. An approach based on a clustering method, i.e., generalized fuzzy algorithmic scheme (GFAS), is introduced to solve the data association problem with a single snapshot. First, the measured data are formed as a matrix. Then, a clustering method is used for data association. This clustering is a procedure of collecting information of each target from the measurements of multiple sensors. Thus, the results of clustering will be used to associate the measured data with their originated targets. In order to overcome the problem that GFAS may associate more than one piece of data measured by the same sensor with a same target, an improved GFAS (IGFAS) is proposed. Different from other methods for data association, only one snapshot is required in the proposed method. Furthermore, the computational complexity will not increase exponentially as the number of targets or sensors increase. Simulation results validate the effectiveness of the proposed method.