Hang Woon Lee
(Advisor: Prof. Koki Ho]

will propose a doctoral thesis entitled,

Design and Operations of Satellite Constellations for Complex Regional Coverage

On

Thursday, December 2 at 1 p.m.

BlueJeans: https://bluejeans.com/681378614/2781

Abstract
Regional coverage satellite constellations (regional constellations, in short) are considered attractive business solutions to circumvent technical, economic, and regulatory issues associated with global constellation systems. However, the use of these systems has been generally restricted to geostationary/synchronous (GSO) orbits. While non-GSO systems are deemed to achieve better performance for mission-critical attributes such as imaging resolution, latency, and launch cost, the design of such systems possesses several technical challenges. Motivated by this, this work aims to address technical challenges that reside in the design and operations of regional constellations. Specifically, we wish to address the following open research questions: 1) How do we optimally design a satellite constellation for a region of interest? What does it mean to be a regional constellation in a non-geostationary orbital regime? 2) How do we optimally reconfigure a satellite constellation for a new mission profile or in case of satellite failure?

The first contribution of this work is the establishment of a novel discrete-time, periodic, and finite-state automaton model that relates the configuration of a satellite constellation with its coverage state with respect to a set of target areas. Based on this model, we formulate the regional constellation design problem (RCDP) as an integer linear program that enables a system designer to optimize the design of a satellite constellation pattern for complex regional coverage. Its unique formulation as a subclass of set covering problems allows system designers to leverage various established exact and heuristic methods developed in other domains (e.g., cyclic staffing personnel problem). The second contribution of this work is the establishment of an integrated optimization framework that models and optimizes the operations of satellite constellations under various mission operational variations. We address the limitations of the state-of-the-art by integrating the constellation design problem and the constellation transfer problem, which are considered independent and serial in nature. To this end, we propose a particular constellation design problem formulation called the maximum coverage problem to overcome the inherent limitation of RCDP by considering the realistic settings of the constellation reconfiguration. The integrated bi-objective constellation design and transfer model enjoys the benefits provided by integer linear programming. However, it suffers from the curse of dimensionality for large instances. Motivated by this challenge, this work proposes a computationally-efficient heuristic method based on the epsilon-constraint reformulation and the Lagrangian relaxation technique. We develop a randomized local search method based on our definition of the combinatorial 1-exchange neighborhood. Numerical results attest to the near-optimality of the obtained heuristic solutions. Finally, we propose an extension of the optimal reconfiguration framework by integrating features such as the timely responsiveness constraint, multi-stage reconfiguration, and non-repeating ground track orbits.

Committee

• Prof. Koki Ho – School of Aerospace Engineering (advisor)
• Prof. Brian Gunter – School of Aerospace Engineering
• Prof. E. Glenn Lightsey – School of Aerospace Engineering