Efficient and Robust Cluster Identification for Ultra-Wideband Propagations Inspired by Biological Ant Colony Clustering

Bin Li*, Chenglin Zhao, Haijun Zhang, Zheng Zhou, Arumugam Nallanathan

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

14 Citations (Scopus)


Cluster identification of ultra-wideband (UWB) propagations is of great significance to the parameter extraction and measurement analysis of channel modeling. In this paper, we address this challenging problem within a promising biological processing framework. Both the two large-scale characteristics of each multipath component, i.e., the decaying amplitude and the time of arrivals, are organically combined and fully explored in the suggested cluster identification algorithm. Each resolvable trajectory component is first projected onto a 2-D amplitude-time plane and further modeled as a virtual ant-agent, which can move around in this 2-D workspace with a preference to the high local-environment similarity. By establishing a subtle population similarity and specifying an efficient position adaptation strategy, cluster identifications can be realized by the biological ant colony clustering procedure. Owing to the population-based intelligence and the involved positive-feedback collaboration during the agents evolution, the suggested algorithm can efficiently identify the involved multiple clusters in a completely automatic manner. Experiments on UWB channels validate the proposed method. The practical parameter configuration is analyzed, and a group of numerical performance metrics is derived. As demonstrated by numerical investigations, multiple clusters involved in UWB channel impulse responses can be accurately extracted.

Original languageEnglish
Pages (from-to)286-300
Number of pages15
JournalIEEE Transactions on Communications
Issue number1
Publication statusPublished - 14 Jan 2015


  • Ultra-wideband propagations
  • cluster identification
  • ant colony clustering
  • population similarity


Dive into the research topics of 'Efficient and Robust Cluster Identification for Ultra-Wideband Propagations Inspired by Biological Ant Colony Clustering'. Together they form a unique fingerprint.

Cite this