9th International Conference on Cognitive Radio Oriented Wireless Networks

Research Article

A Study on the Design Space for Harm Claim Thresholds

Download433 downloads
  • @INPROCEEDINGS{10.4108/icst.crowncom.2014.255404,
        author={Janne Riihij\aa{}rvi and Andreas Achtzehn and Petri M\aa{}h\o{}nen and Pierre de Vries},
        title={A Study on the Design Space for Harm Claim Thresholds},
        proceedings={9th International Conference on Cognitive Radio Oriented Wireless Networks},
        keywords={harm claim thresholds interference limits order statistics radio regulation drive tests measurements},
  • Janne Riihijärvi
    Andreas Achtzehn
    Petri Mähönen
    Pierre de Vries
    Year: 2014
    A Study on the Design Space for Harm Claim Thresholds
    DOI: 10.4108/icst.crowncom.2014.255404
Janne Riihijärvi1,*, Andreas Achtzehn1, Petri Mähönen1, Pierre de Vries2
  • 1: RWTH Aachen University
  • 2: University of Colorado
*Contact email: jar@inets.rwth-aachen.de


Harm claim threshold policies will offer more flexibility in radio regulations. By defining bounds on the acceptable field strength over time and space, these policies will introduce quantitative measures to inter-system coexistence questions. However, this requires tangible and sufficient means of proving claims of interference through spectrum measurements. In this paper we study drive test requirements in terms of necessary sample set sizes and decorrelated sampling. We present order statistics and extreme value theory as powerful mathematical tools for describing necessary confidence intervals, and test their practical viability with data from an extensive measurement campaign. Furthermore, we discuss by means of example necessary extensions to planning tools for wireless networks operating under an interference threshold policy. Our results emphasize that future regulations will need to be accompanied by a rigorous specification of evidence collection requirements in order to compensate for bias and correlation structure in the spectrum data.