Satellite Constellation Pattern Optimization for Complex Regional Coverage

TL;DR

This paper presents a novel optimization approach for designing satellite constellations tailored to complex regional coverage, outperforming traditional methods especially in time-varying and spatially-varying scenarios.

Contribution

It introduces a formalized circular convolution framework and a binary integer linear programming method for optimal regional satellite constellation design.

Findings

The new approach achieves optimal constellation patterns in various scenarios.

The binary integer linear programming method outperforms baseline symmetry-based designs.

Five examples demonstrate the effectiveness of the proposed optimization techniques.

Abstract

The use of regional coverage satellite constellations is on the rise, urging the need for an optimal constellation design method for complex regional coverage. Traditional constellations are often designed for continuous global coverage, and the few existing regional constellation design methods lead to suboptimal solutions for periodically time-varying or spatially-varying regional coverage requirements. This paper introduces a new general approach to design an optimal constellation pattern that satisfies such complex regional coverage requirements. To this end, the circular convolution nature of the repeating ground track orbit and common ground track constellation is formalized. This formulation enables a scalable constellation pattern analysis for multiple target areas and with multiple sub-constellations. The formalized circular convolution relationship is first used to derive a baseline constellation pattern design method with the conventional assumption of symmetry. Next, a novel method based on binary integer linear programming is developed, which aims to optimally design a constellation pattern with the minimum number of satellites. This binary integer linear programming method is shown to achieve optimal constellation patterns for general problem settings that the baseline method cannot achieve. Five illustrative examples are analyzed to demonstrate the value of the proposed new approach.