Rapid Uncertainty Propagation and Chance-Constrained Path Planning for Small Unmanned Aerial Vehicles

TL;DR

This paper presents a fast method for propagating uncertainty and planning safe paths for small unmanned aerial vehicles using linear covariance propagation and a quadratic programming collision detection, integrated with an informed RRT* algorithm.

Contribution

It introduces a novel combination of linear covariance propagation with quadratic programming collision detection and an informed RRT* for efficient chance-constrained path planning.

Findings

Rapid validation of flight plans for sUAS models.

Efficient chance-constrained path planning demonstrated.

Applicable to fixed-wing and quadrotor sUAS.

Abstract

With the number of small Unmanned Aircraft Systems (sUAS) in the national airspace projected to increase in the next few years, there is growing interest in a traffic management system capable of handling the demands of this aviation sector. It is expected that such a system will involve trajectory prediction, uncertainty propagation, and path planning algorithms. In this work, we use linear covariance propagation in combination with a quadratic programming-based collision detection algorithm to rapidly validate declared flight plans. Additionally, these algorithms are combined with a Dynamic, Informed RRT* algorithm, resulting in a computationally efficient algorithm for chance-constrained path planning. Detailed numerical examples for both fixed-wing and quadrotor sUAS models are presented.