Simultaneous adaptive localization of a wireless sensor network

  • Authors:
  • Koushil Sreenath;Frank L. Lewis;Dan O. Popa

  • Affiliations:
  • University of Texas at Arlington, Arlington, Texas;University of Texas at Arlington, Arlington, Texas;University of Texas at Arlington, Arlington, Texas

  • Venue:
  • ACM SIGMOBILE Mobile Computing and Communications Review
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

A range-free approach for adaptive localization of un-localized sensor nodes employing a mobile robot with GPS is detailed. A mobile robot navigates through the sensor deployment area broadcasting its positional estimate and the uncertainty in its estimate. Distributed computationally-inexpensive, discrete-time Kalman Filters, implemented on each static sensor node, fuse information obtained over time from the robot to decrease the uncertainty in each node's location estimate. On the other hand, due to dead reckoning and other systematic errors, the robot loses positional accuracy over time. Updates from GPS and from the localized sensor nodes serve in improving the localization uncertainty of the robot. A Continuous-Discrete Extended Kalman Filter (CD EKF) running on the mobile robot fuses information from multiple distinct sources (GPS, various sensors nodes) for robot navigation. This two-part procedure achieves simultaneous localization of the sensor nodes and the mobile robot. Also presented is an adaptive localization strategy to navigate the mobile robot to the area of least localized sensor nodes. This ensures that the robot maneuvers to an area where the sensor nodes possess the largest uncertainty in location, so that it can maximize the usefulness of its positional information in best localizing the overall network.