Hu, Qiang (2005-12). Robotic localization of hostile networked radio sources with a directional antenna. Master's Thesis. Thesis uri icon

abstract

  • One of the distinguishing characteristics of hostile networked radio sources (e.g., enemy sensor network nodes), is that only physical layer information and limited medium access control (MAC) layer information of the network is observable. We propose a scheme to localize hostile networked radio sources based on the radio signal strength and communication protocol pattern analysis using a mobile robot with a directional antenna. We integrate a Particle Filter algorithm with a new sensing model which is built on a directional antenna model and Carrier Sense Multiple Access (CSMA)-based MAC protocol model. we model and analyze the channel idle probability and busy collision probability as a function of the number of radio sources in the CSMA protocol modeling. Based on the sensing model, we propose a particle-filter-based scheme to simultaneously estimate the number and the positions of networked radio sources. We provide a localization scheme based on the method of steepest descent for the purpose of performance comparison. Simulation results demonstrate that our proposed localization scheme has a better success rate than the scheme based on the steepest descent at different tolerant distances.
  • One of the distinguishing characteristics of hostile networked radio sources (e.g.,
    enemy sensor network nodes), is that only physical layer information and limited
    medium access control (MAC) layer information of the network is observable. We
    propose a scheme to localize hostile networked radio sources based on the radio signal
    strength and communication protocol pattern analysis using a mobile robot with a
    directional antenna. We integrate a Particle Filter algorithm with a new sensing
    model which is built on a directional antenna model and Carrier Sense Multiple
    Access (CSMA)-based MAC protocol model. we model and analyze the channel
    idle probability and busy collision probability as a function of the number of radio
    sources in the CSMA protocol modeling. Based on the sensing model, we propose a
    particle-filter-based scheme to simultaneously estimate the number and the positions
    of networked radio sources. We provide a localization scheme based on the method
    of steepest descent for the purpose of performance comparison. Simulation results
    demonstrate that our proposed localization scheme has a better success rate than the
    scheme based on the steepest descent at different tolerant distances.

publication date

  • December 2005