Low-thrust trajectory optimization for the solar system pony express

Highlights

• Data mule spacecraft flying on Earth–Mars cycler orbits can be used to enhance the capabilities of the Deep Space Network (DSN).

• Earth–Mars cycler orbits computed in a patched-conics model are a good initial guess to compute similar cycler orbits in the ephemeris model.

• Low-thrust propulsion can be leveraged to design fuel-optimal cycler trajectories.

• The propellant mass required to utilize the Earth–Mars cycler orbits is compatible with ESPA-class satellites ($<$ 500 kg).

Abstract

We present a method for designing and optimizing Earth–Mars cycler orbits for the Solar System Pony Express (SSPE) mission concept. The SSPE is aimed at augmenting the capabilities of the Deep Space Network (DSN) by enabling high latency and high bandwidth communications through the use of interplanetary data mules. The SSPE will leverage a network of smallsats, outfitted with optical laser communications, and by exploiting cycler orbits it will perform flybys of Mars at least once a year, retrieving 1–3 petabits of data per flyby from missions already operating at the planet. This data will then be downlinked to Earth during the next Earth flyby, also occurring about once a year. The SSPE satellites are launched as a rideshare payload with another Mars-bound mission. Each data mule flying in the SSPE network is equipped with a low-thrust propulsion system that is used for cycler orbit injection and for targeting flybys of Earth and Mars. Cycler orbits are first computed using a patched conic approximation and impulsive thrust. These candidate solutions are then transitioned into a higher-fidelity ephemeris model, and using an indirect optimization method, fuel-optimal low-thrust transfers are computed between each pair of flybys. We present three high-fidelity candidate solutions where a 500 kg data mule is inserted into an Earth–Mars cycler orbit using the NASA Evolutionary Xenon Thruster (NEXT). In all three cases, the results show that the fuel consumption is feasible from a mission design perspective and that solutions obtained in the patched conics model can be successfully transitioned to the ephemeris model using low-thrust propulsion.

Introduction

Telecommunications is a fundamental aspect of space exploration. Satellites, interplanetary spacecraft, landers and rovers are crucial assets designed to acquire data and transmit it back to Earth, where thorough analysis and investigation helps further our collective knowledge of the Solar System and the Universe. Often, the impact of a scientific mission is assessed based on the amount of data it returned over its lifetime and the number of publications that made use of the data. The data volumes generated by a mission depend on several factors, such as the design of the spacecraft, the technology used for telecommunication, and the distance from Earth.

Currently, NASA supports interplanetary spacecraft missions using the Deep Space Network (DSN). The DSN is a global radio telecommunication network composed of three ground complexes, spaced apart by about 120${}^{\circ }$ around the Earth, and each facility operates several 34-m antennas and one 70-m antenna. Thanks to the strategic placement of the facilities and the large dish antennas, the DSN is able to constantly communicate with spacecraft over interplanetary distances. However, the use of radio waves and the large distances between the antenna and the probes mean that the links are power-starved, which restricts the maximum data volume, and as a result restricts the maximum data volume that a mission can return. In addition, the DSN is a resource that is shared by many missions which further reduces the data volume that can be returned by a single probe.

The Solar System Pony Express (SSPE) [1] is a mission concept that is being investigated at the Jet Propulsion Laboratory as part of the NASA Innovative Advanced Concepts (NIAC) program. The objective of the study is to augment the data transmission rates of the DSN using the idea of data mules. Data mules are small spacecraft that travel to a remote location (e.g., Mars) where they acquire data in close range to the probe’s transmitter and then carry the data back to Earth where it is downlinked in close range to the receiver. This enables high latency and high bandwidth communication. In the case of a space network, the data transmission rates can be further increased by utilizing optical telecommunications (e.g., $>1$ Gbps have been demonstrated in an optical link between the Moon and Earth, a value at least one order of magnitude larger than with RF communication systems), which allows for a larger bandwidth but also requires strict pointing accuracy [2].

A network of interplanetary data mules could be established by exploiting cycler orbits [3]. These are trajectories that regularly encounter two or more celestial bodies along their path and require a modest amount of propellant for trajectory correction maneuvers. For the purpose of this study we simulate a network of Earth–Mars cyclers, but the concept could be extended to Venus [4] or the outer planets [5]. In the current mission scenario, one or more smallsats could be launched as a secondary payload onboard a mission bound for Mars. A simplified diagram of the mission concept is shown in Fig. 1. After launch, the data mules use their own low-thrust propulsion system to inject into a cycler orbit and target subsequent flybys of Earth and Mars. During the Mars flybys, data is uplinked from spacecraft already operating at Mars (in orbit or on the surface) and during Earth flybys data is downlinked back to Earth.

In this paper we focus specifically on the trajectory design and optimization of Earth–Mars cycler orbits for the SSPE mission. We simulate trajectories that make use of low-thrust propulsion and include a high-fidelity model that incorporates the gravity of the Sun, Earth and Mars. Low-thrust space missions are becoming more common due to the benefits afforded by ion engines, which are more efficient than chemical engines due to their high specific impulse. They are also much smaller/lighter which allows for the design of smaller spacecraft that can be launched economically as a secondary payload.

Two main classes of methods are used to solve for low-thrust spacecraft trajectories. The first class, direct methods, transcribes the system dynamics into a nonlinear programming problem. Generally, direct methods are robust to an initial guess, but are computationally expensive and are not guaranteed to result in an optimal solution. The second class, indirect methods, seeks to satisfy the necessary conditions for optimality. Although indirect methods ensure that the solution is at least locally optimal, they are extremely sensitive to initial guesses and convergence can be difficult. To overcome this challenge, we apply hyperbolic tangential smoothing to the thrust profile and generate the optimal bang–bang thrust profile by continuation [6], [7], which improves convergence significantly. The indirect method with continuation has proven to be very reliable and is the method of choice for this paper.