Localization of Chemical Sources Using Stochastic Differential Equations in Realistic Environments

A. Mohammed, and A. Jeremic
McMaster University, Hamilton, L8S4K1, Canada

Signal processing algorithms for chemical sensing/monitoring have been subject of considerable research interest in the recent years mainly due to their diverse applicability. When the concentration of chemical agent is small, the dispersion of particles is governed by stochastic differential equations describing more complex motion mechanisms such as Brownian motion.

In this paper we propose the computational framework for solving estimation problems using stochastic differential equations and finite-element analysis environments. We demonstrate that the location of the source can be estimated more accurately by accounting for the stochastic nature of dispersion which is unaccounted for in classically used techniques based on Fick’s law.