Sensor Network for Odor Source Localization

Introduction

Odor source localization is the problem of finding the location of an odor source in the environment. It is natural problem in biological world. Many life forms rely on the capability of tackling this problem for survival. Also this problem arises in many circumstances in human life such as in disaster situation where many entrapped victims under collapsed building could be found by tracking their odors. We have been so far relying on dog’s sensing capability to track down an odor source. Training a dog to search for an odor source under human command is laborious and expensive work. Moreover tracking dog performance may be weakened due to physical and mental exhaustion. Thus it is of interest to develop an inexpensive yet reliable and robust method to tackle odor source localization problem.

Objective

This research aims to develop a technique and strategy to locate odor source using sensor network technology. Several sensing nodes work collaboratively to improve localization performance in term of searching speed compared to when using single sensing node.

Methodology

The obvious advantages of using multiple sensing nodes for odor source localization are the capability to search over a wider area in shorter time compared to when using single node. The nodes collaborate with each other to search for the odor concentration and to locate the odor source location. In this coordination, they share information on odor concentration, their location and collectively estimate the source location.

The task of odor source localization in general can be divided into three sub-task:

  1. Searching for odor plume

  2. Following odor plume up to the source

  3. Declaration of odor source

There are various strategies of source localization. Those strategies can be classified into strategies learned from behavior of living organisms / animal (biologically-inspired) and those of developed based on engineering principles. Localization using sensor network falls into artificial strategy since it harness capabilities that are not possessed by living organisms such as sharing information and collective discerning on gathered information to establish probabilistic map for odor source location.