Airborne Plume Localization and Tracking

摘要

In this paper we consider the problem of localization and tracking of an airborne instantaneous plume released from a point source using concentration measurements made by a network of sensors. Using the Gaussian plume model and stochastic modeling of the wind velocity, we formulate plume localization as a nonlinear least squares (LS) problem. We then propose an efficient method for its solution via explicit linear LS formulas rather than by using computationally involved numerical methods.In addition, the uncertainty regarding the diffusion coefficients is modeled using a multiple model (MM)method. Results from a simulation study conducted with realistic plume propagation in a homogeneous wind field demonstrate the feasibility and efficiency of the proposed linear LS solution. Significant improvement is achieved through the use of multiple model method.