Multisensor Integration for Declaring the Odor Source of a Plume in Turbulent Fluid-Advected Environments

Abstract

A mission of chemical plume tracing (CPT) is to navigate an autonomous underwater vehicle (AUV) to find a chemical plume, to trace the plume to its source, and to declare the source location. This paper presents a multisensor integration strategy for source declaration subsumed in a behavior based control architecture with several individual behaviors. The performance of the proposed strategy is evaluated using a simulated turbulent fluid environment. The results show that, for 1000 test runs, the proposed strategy reaches a success rate in declaring the odor source of 98% and an average error of the declared source location about 3.27 meters in an operation area with length scales of 100 meter. Source verification is developed using a fuzzy reasoning based segmentation algorithm to recognize the odor source in color images acquired by a visual sensor in the vicinity of the odor source.