Biologically-Inspired Search Algorithms for Locating Unseen Odor Sources

Abstract

Many animals, ranging from bacteria, through nematodes, to insects, fish and mammals, use air- or water-borne plumes of odor molecules to locate distant unseen resources. We have used our knowledge of the well studied pheromone plume tracking behavior of male moths, as the foundation for simulation and robotic experiments aimed at learning more about the mechanisms underlying biological systems. We extended this knowledge to engineering approaches to generate an autonomous artificial agent that is able to track chemical plumes through complex environments. We have developed simulation models and robotic control systems that adapt their performances to the local odor and wind environment and perform more successfully than those that do not adapt. Algorithms that use Bayesian estimation adapt their performances to the width of the odor plume and 'home-in' on the odor source in a manner similar to the behavior of flying moths tracking odor plumes. Mobile robot odor tracking experiments show that simple steering algorithms are more successful at locating the odor source than those with internal models of the wind sensing system.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 20, 2002
Accession Number
ADA402125

Entities

People

  • Mark A Willis

Tags

Communities of Interest

  • Autonomy

DTIC Thesaurus Topics

  • Algorithms
  • Computer Simulations
  • Control Surfaces
  • Control Systems
  • Data Sets
  • Environment
  • Navigation
  • Pheromones
  • Simulations
  • Statistics
  • Steering
  • Systems Biology
  • Wind
  • Wind Direction
  • Wind Tunnels

Readers

  • Aerospace Propulsion Engineering.
  • Neural Network Machine Learning.
  • Robotics and Automation.

Technology Areas

  • AI & ML
  • AI & ML - Autonomous Systems
  • Autonomy
  • Autonomy - Autonomous System Control