查看文章

Clustering phenomenon exists in the Ultra-Wideband (UWB) impulse responses. Although it is feasible to manually identify clusters via visual inspection, this task becomes very difficult and time consuming when a large amount of data needs to be processed. Furthermore, visual inspection highly depends on the person who performs the cluster identification task, which may cause inconsistent or unrepeatable results. In this paper, we propose an automatic procedure to identify clusters in UWB impulse responses. The algorithm takes both the amplitude and time delay into account when finding the clusters. The algorithm performs well. If the threshold is set correctly, there will not be any error.