Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Optimal Ground Station Selection for Low-Earth Orbiting Satellites

Duncan Eddy, Michelle Ho, Mykel J. Kochenderfer

TL;DR

The paper addresses optimal ground-station provider and location selection for low-Earth-orbit missions under a GSaaS model. It formulates the problem as an integer program that selects providers, stations, and contact opportunities to optimize mission performance under cost and timing constraints, augmented by a surrogate optimization approach that evaluates performance over a reduced time window $T_{sim}$ while targeting a longer mission horizon $T_{opt}$. The study implements three objective functions—minimize total cost, maximize data downlink, and minimize the maximum contact gap—along with a versatile set of constraints that link contacts to stations and providers, enforce scheduling limits, and cap costs. Experimental results across random scenarios and constellation sizes show that an IP-optimized network can outperform fixed one- or two-provider configurations in both cost efficiency and data throughput, with runtimes that remain practical for planning across sizeable satellite fleets. These findings support GSaaS adoption for scalable, data-intensive missions and highlight avenues for future work in multi-objective optimization and unconstrained terrestrial placement, with open-source tooling available to the community.

Abstract

This paper presents a solution to the problem of optimal ground station selection for low-Earth orbiting (LEO) space missions that enables mission operators to precisely design their ground segment performance and costs. Space mission operators are increasingly turning to Ground-Station-as-a-Service (GSaaS) providers to supply the terrestrial communications segment to reduce costs and increase network size. However, this approach leads to a new challenge of selecting the optimal service providers and station locations for a given mission. We consider the problem of ground station selection as an optimization problem and present a general solution framework that allows mission designers to set their overall optimization objective and constrain key mission performance variables such as total data downlink, total mission cost, recurring operational cost, and maximum communications time-gap. We solve the problem using integer programming (IP). To address computational scaling challenges, we introduce a surrogate optimization approach where the optimal station selection is determined based on solving the problem over a reduced time domain. Two different IP formulations are evaluated using randomized selections of LEO satellites of varying constellation sizes. We consider the networks of the commercial GSaaS providers Atlas Space Operations, Amazon Web Services (AWS) Ground Station, Azure Orbital Ground Station, Kongsberg Satellite Services (KSAT), Leaf Space, and Viasat Real-Time Earth. We compare our results against standard operational practices of integrating with one or two primary ground station providers.

Optimal Ground Station Selection for Low-Earth Orbiting Satellites

TL;DR

The paper addresses optimal ground-station provider and location selection for low-Earth-orbit missions under a GSaaS model. It formulates the problem as an integer program that selects providers, stations, and contact opportunities to optimize mission performance under cost and timing constraints, augmented by a surrogate optimization approach that evaluates performance over a reduced time window while targeting a longer mission horizon . The study implements three objective functions—minimize total cost, maximize data downlink, and minimize the maximum contact gap—along with a versatile set of constraints that link contacts to stations and providers, enforce scheduling limits, and cap costs. Experimental results across random scenarios and constellation sizes show that an IP-optimized network can outperform fixed one- or two-provider configurations in both cost efficiency and data throughput, with runtimes that remain practical for planning across sizeable satellite fleets. These findings support GSaaS adoption for scalable, data-intensive missions and highlight avenues for future work in multi-objective optimization and unconstrained terrestrial placement, with open-source tooling available to the community.

Abstract

This paper presents a solution to the problem of optimal ground station selection for low-Earth orbiting (LEO) space missions that enables mission operators to precisely design their ground segment performance and costs. Space mission operators are increasingly turning to Ground-Station-as-a-Service (GSaaS) providers to supply the terrestrial communications segment to reduce costs and increase network size. However, this approach leads to a new challenge of selecting the optimal service providers and station locations for a given mission. We consider the problem of ground station selection as an optimization problem and present a general solution framework that allows mission designers to set their overall optimization objective and constrain key mission performance variables such as total data downlink, total mission cost, recurring operational cost, and maximum communications time-gap. We solve the problem using integer programming (IP). To address computational scaling challenges, we introduce a surrogate optimization approach where the optimal station selection is determined based on solving the problem over a reduced time domain. Two different IP formulations are evaluated using randomized selections of LEO satellites of varying constellation sizes. We consider the networks of the commercial GSaaS providers Atlas Space Operations, Amazon Web Services (AWS) Ground Station, Azure Orbital Ground Station, Kongsberg Satellite Services (KSAT), Leaf Space, and Viasat Real-Time Earth. We compare our results against standard operational practices of integrating with one or two primary ground station providers.

Paper Structure

This paper contains 14 sections, 25 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Problem Formulation
  3. IP Formulation
  4. Objective Functions
  5. Constraint Functions
  6. Experiments
  7. Mission Cost Minimization Problem
  8. Mission Data Maximization Problem
  9. Conclusions
  10. Commercial GSaaS Station List
  11. Surrogate Optimization Assumption Analysis
  12. Mission Cost Minimization Problem
  13. Data Volume Maximization Problem
  14. Maximum Communications Gap Minimization Problem

Figures (4)

  • Figure 1: Plot of all currently operational ground stations and associated communications cones for Atlas (purple), AWS (orange), Azure (blue), KSAT (red), Leaf Space (green), and Viasat (yellow). Communications cones assume 525 altitude and $10^{\circ}$ minimum elevation angle. See Appendix \ref{['sec:appendix_station_list']} for station list.
  • Figure 2: Performance comparisons of IP-optimized ground station network with one- and two-provider solutions. Normalized total cost (left), normalized total data downlink (middle) and average optimizer runtime (right). The normalized value is the value of a solution divided by the value of the optimal solution to that scenario. Closer to 1 is better for all normalized values.
  • Figure 3: Performance comparisons of IP-optimized ground station network with one- and two-provider solutions. Normalized total data downlink (left), normalized total cost (middle) and average optimizer runtime (right). The normalized value is the value of a solution divided by the value of the optimal solution to that scenario. Closer to 1 is better for the objective value and worse for cost.
  • Figure 4: Contact simulation statistics for 100 satellites propagated over 1, 2, 3, 5, 7, 10, 20, 30, 50, 60, 90, 100, and 180-day simulation windows. Mean contact gap duration (left), mean contact duration (middle) and mean number of contacts (right).