Project title: Optimal Sampling Strategy for Advection-Diffusion Parameter Estimation using Vehicles with Nonholomonic Dynamics


Andrew White

Abstract

There exist many physical transport phenomenon that are governed by advection-diffusion processes. Some of these processes occur naturally due to local climate conditions, such as the growth and spread of harmful algae blooms. Others occur without natural cause, such as the release of toxic chemicals. Due to the harmful nature of some of these processes, researchers have been developing methods to track the source of such advection-diffusion processes. In this paper, we present an optimal sensing strategy for a group of mobile sensing agents to estimate the parameters of an advection-diffusion process with an impulse chemical source. Our approach to solving this problem is to use a nonlinear least-square algorithm on the advection-diffusion model over a spatially averaged velocity field. The advection-diffusion process parameters will be averaged over a control volume. The control volume will contain the sensing agents. To control the position and spacing between each of the sensing agents a consensus algorithm will be used. The concept will be tested using two-wheeled vehicles with nonholonomic dynamics. Nonlinear control synthesis techniques will be used to stabilize the nonholonomic dynamics of the two-wheeled vehicles. Comparisons of performance between the methods used will be made.



pdf file