Application of Nonparametric Belief Propagation for Localization

Project Type: Student Project
Student: Ricard Comas Martinez

Short Description

A fundamental task in a wireless sensor network is the localization of the sensor nodes. Typically, the sensors measure the pair-wise distances between them. Based on this set of measurements, the location of all the sensors is estimated. This is preferably done with statistical methods which can, for example, not only provide the node locations, but also the respective uncertainties in the form of a posterior distribution. For this, graphical models are often used.

However, the implementation of these concepts can be tricky. Recently, nonparametric belief propagation (NBP) has been proposed as a possible solution. NBP can be seen as a generalization of particle filtering and allows the sensor localization problem to be solved  without restrictions to e.g. Gaussian measurement models.

The aim of this project is a reference implementation of NBP for localization in wireless sensor networks.  

Your Tasks

  • Literature review on graphical models and nonparametric belief propagation.
  • Implementation of NBP in Matlab

Your Profile

  • Interest in wireless sensor networks and localization
  • Some basic Matlab skills


Thomas Buchgraber, Paul Meissner, Phone: 0136/873 4386 or 4363


[1] Ihler, A., Fisher, J., Moses, R. and Willsky, A.: Nonparametric Belief Propagation for Self-Localization of Sensor Networks, IEEE Journal on Selected Areas in Communications, 2005