Robot Navigation in a Decentralized Landmark-Free Sensor Network

  • Authors:

  • Travis Mercker;Maruthi Akella;Jorge Alvarez

  • Affiliations:

  • Department of Aerospace Engineering and Engineering Mechanics, University of Texas at Austin 1 University, Austin, USA 78712;Department of Aerospace Engineering and Engineering Mechanics, University of Texas at Austin 1 University, Austin, USA 78712;Department of Aerospace Engineering and Engineering Mechanics, University of Texas at Austin 1 University, Austin, USA 78712

  • Venue:
  • Journal of Intelligent and Robotic Systems
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

A wireless sensor network has the ability to autonomously perform event detection over large areas, but power and/or cost constraints limit the addition of equipment such as cameras onto sensor modules to verify events. Accordingly, verification must be performed by an independent mobile robotic vehicle which has sensing equipment for improved event detection. The main challenge, however, is that the robotic vehicle itself is typically located somewhere in the sensor field and has no prior knowledge of the geographic location of the event. In this paper, we specifically focus upon the scenario of navigating a robotic vehicle through a stationary wireless sensor network as a means to perform event verification. The underlying assumptions are that the robotic vehicle has distance traveled and heading measurements, but the only additional information provided by the stationary sensors is a communication boundary. More significantly, we emphasize that under this scenario, neither the robot nor the ground-fixed sensing nodes have location or any other geographical landmark information. The paper introduces two distinct and novel navigation algorithms that permit the robotic vehicle to travel from one fixed node to another along a communication path established in an ad-hoc fashion. These navigation algorithms have been tested on a newly developed UTrekr robotic vehicle within a hardware based ground-fixed sensor network and under assumption of perfect communication and network operations, we report a nearly 100% success rate even while using open-loop robot control.