Model-agnostic Methodology for High-Altitude Balloon Mission Planning

Abstract

High-altitude balloons (HABs) are increasingly used for atmospheric science, technology validation, Earth observation, and resilient communications. There is a need for modeling HABs in multi-domain missions, and a key challenge for incorporating them is their flight trajectory dependence on atmospheric conditions. The difficulty in working with atmospheric conditions is that forecasts can only be relatively accurate to a specific date, a few days in advance. These atmospheric conditions, such as wind fields, vary between days, months, and years. Therefore, planning a date-agnostic multi-platform mission with HAB vehicle dynamics that are date dependent is an unsolved problem. There is a need for modeling high-altitude balloons in this way for multi-domain mission planning, and with technology getting smaller, specifically a need to model smaller latex high-altitude balloons. The goal of this research is to develop a method to incorporate wind-dependent balloon trajectories into mission planning tools without known atmospheric conditions. This dissertation proposes a novel model-agnostic set of probabilistic state equations for floating latex high-altitude balloon flights. Individual latex flights have been proven to closely match with the two commonly used sets of latex balloon dynamics, but a large scale analysis of simulation vertical and horizontal motion error to actual balloon flight trajectories has not been published. This dissertation presents the error quantification of two sets of latex high-altitude balloon vertical dynamics, a 2nd-order and a 1st-order model. The models are found to have similar average horizontal errors within 1.5km of each other and average descent vertical error within 1km. The ascent vertical error for the 1st-order model, however, has twice the average error as the 2nd-order, but still less than 650m on average. Using the 1st-order model, 700,000+ float flight trajectories were generated and analyzed to inform a set of probabilistic novel model-agnostic state equations. To demonstrate the model-agnostic state equations, a metaheuristic optimization mission planning tool was applied to evaluate the resiliency of a degraded LEO Walker-Delta satellite constellation. One study explores the trends in Walker-Delta design parameters that impact constellation coverage degradation to understand the worst-case degradation scenario. A second study then evaluates the ability for latex high-altitude balloons to reconstitute a degraded LEO Walker-Delta satellite constellation. For a three-day scenario observing the state of Virginia with a satellite constellation of 60% coverage and 70 minutes average maximum revisit time, launching 64 balloons sequentially from six locations improves the region coverage to 92% with a 44-minute maximum revisit time. With less than 90 balloons, the coverage increases to almost 95% with 37-minute maximum revisit time.