Plume estimation using static and dynamic formations of unmanned aerial vehicles

Abstract

This paper presents a model based approach for the real-time estimation of a gaseous plume using a time-varying formation of unmanned aerial vehicles (UAVs). Such a dynamic formation is necessitated by sensor technology limitations that can only obtain concentration measurements and not spatial gradients of concentration. It is assumed that gas is released in the atmosphere by a stationary or a moving aerial source. Each UAV in the flying formation has a single concentration sensor onboard. The estimation scheme, which uses the concentration measurements provided by the UAVs, is built in the form of a Luenberger observer. The plume is reconstructed by solving numerically, and in real-time, the advection-diffusion equation using the finite volume method on a dynamically adapted computational grid. A leader-follower approach is used for the control scheme of the formation flight in order to provide approximations of the spatial gradients. The control signals for the leader, which are provided by the Lyapunov-redesign method, incorporate the concentration gradient measurements. The latter is approximated by simultaneous concentration measurements from the followers. The follower-UAVs keep physics-based prescribed distances from the leader, which are dictated by the performance of the estimator and must follow the estimated length scale for meaningful approximation of spatial gradients of the plume concentration. The numerical simulation results demonstrate the performance of the proposed approach.