2006
DOI: 10.1049/el:20063917
|View full text |Cite
|
Sign up to set email alerts
|

Improving clustering performance using multipath component distance

Abstract: The problem of identifying clusters from MIMO measurement data is addressed. Conventionally, visual inspection has been used for cluster identification, but this approach is impractical for a large amount of measurement data. For automatic clustering, the multipath component distance (MCD) is used to calculate the distance between individual multipath components estimated by a channel parameter estimator, such as SAGE. This distance is implemented in the well-known KMeans clustering algorithm. To demonstrate t… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
27
0

Year Published

2006
2006
2024
2024

Publication Types

Select...
4
3
1

Relationship

0
8

Authors

Journals

citations
Cited by 75 publications
(27 citation statements)
references
References 3 publications
0
27
0
Order By: Relevance
“…MC distance (MCD) [18] including powers of propagation paths is used as a metric for calculating the multidimensional distances between propagation paths and cluster centroids. Propagation paths are set to the clusters based on the shortest MCD.…”
Section: Cluster Parametersmentioning
confidence: 99%
“…MC distance (MCD) [18] including powers of propagation paths is used as a metric for calculating the multidimensional distances between propagation paths and cluster centroids. Propagation paths are set to the clusters based on the shortest MCD.…”
Section: Cluster Parametersmentioning
confidence: 99%
“…Because of simplicity elevation, polarimetric path weight and Doppler are not considered. In order to evaluate the similarity between multipath components and cluster centroids based on the different parameters of the data model, the weighted multipath component distance (MCD) is used [9] [8], , · , .…”
Section: A Data Model and Multipath Component Distance Measurementioning
confidence: 99%
“…A so-called cluster power threshold (CPT) is introduced within the IG, to define the stop criteria based on the minimum power allowed for the smallest cluster. The delay scaling factor (DSF) [9] is introduced to additionally weight the delay in the calculation of the MCD with the IG and clustering. However for fixed CPT and DSF the described approach always converges toward the same centroids, and therefore IG is considered as deterministic procedure.…”
Section: E Deterministic and Random Initialization 1) Initial Guess mentioning
confidence: 99%
See 1 more Smart Citation
“…Scatterer scenario. a recent paper, k-means algorithm with a different distance metric is used to identify clusters from measurement data [11]. From Figs.…”
mentioning
confidence: 99%