With multiple missions planned to target the Moon's south pole, robust and high-performance powered descent guidance algorithms are critical to ensuring safe and autonomous landings. One leading approach is the six degree-of-freedom (6DOF) Successive Convexification (SCVx) algorithm [1], currently utilized in vehicles such as the Falcon 9. SCVx works by iteratively solving a non-convex optimal control problem by approximating it as a sequence of convex subproblems, efficiently generating realtime, constraint optimized trajectories. To further enhance SCVx's robustness and convergence speed, recent work by Briden [2] introduced a transformer-based warm-starting method known as TSCVx. T-SCVx maintains the rigor of SCVx's convex optimization framework while enhancing its responsiveness in dynamic environments, especially where onboard computational resources or time are constrained. In this work, we train a transformer neural network with mission-specific trajectory dispersions to increase safety and convergence reliability of T-SCVx, extending on previous efforts of training a dataset for landing. By providing feasible initial trajectory generations for preliminary targeting activities during the coasting phase of flight prior to powered descent guidance, under performing burns may be effectively accounted for by the trained neural networks, increasing convergence robustness. This methodology of customized neural network training with mission specific dispersions data can be useful for similar initial conditioned/mass properties vehicles landing at or near the same landing site, such as multiple Starship landings. In this work, the simulation is of a transfer from the Artemis Gateway proposed 9:2 Near Rectilinear Halo Orbit (NRHO) around the Moon to a low-lunar oribit (LLO) transitioning to T-SCVx powered terminal descent. During training, the initial trajectory leg between NRHO-LLO is simulated in NASA Goddard's GMAT with an underperforming engine modeled to produce a diverse set of initial conditions for training terminal descent conditions. After the transformer has been trained and tuned, the simulation is run in the Space Teams PRO environment, starting at a variety of terminal descent points derived from the simulated dispersion. This simulation is used to validate the method's ability to maintain accuracy and safety in scenarios with significant trajectory variation. Our mission-informed dispersion-based training outperformed our spherical distribution trained implementation in convergence speed for the case where initial conditions live within both spherical distribution and mission-informed trained transformer (average 8.18 seconds ± 5.16 seconds at $3 \sigma$). The mission-informed trained transformer also shows consistently smaller variance in solve times for all cases while reliably converging on all test cases.

In this paper, we present the Planetary Exploration Transport: a novel concept for a reusable, multi-purpose planetary lander and habitat vehicle. This vehicle is designed to descend from orbit onto a planetary body, support a six-month mission on the surface (including potential relocations), then ascend back to orbit. The vehicle is designed with reusability in mind-the same configuration can be used for the Moon, Mars, and other planets or moons. We propose the use of nuclear thermal rockets for the main thrusters of the vehicle, which allow it to be a single-stage-to-orbit lander on most terrestrial bodies in the solar system. To validate the design, we present simulations of key vehicle systems and example mission scenarios, including descent and landing on the Moon, interactive driving on the Martian surface, and long-term simulations of environmental and life support systems. Lastly, we briefly discuss the outcomes of this project and its future.

In response to the growing need for streamlined hands-free interfaces across diverse fields such as space exploration, education, and industry, we performed a study to determine the most effective method of interacting with Graphical User Interface (GUI) menus in an augmented reality (AR) setting. Focusing on astronauts as the target users, we explored different interaction modalities within an immersive Virtual Reality (VR) environment. Through a suite of user studies, we compared five interaction modalities: voice commands, gaze tracking, eye tracking, popup buttons, and gestures. The goal of this work was to identify the most effective and user-friendly approach to optimize user efficiency and efficacy in hands-free tasks. Participants performed a standardized menu navigation task using each interaction modality. The experiments were designed to simulate scenarios where manual dexterity is restricted, such as during extravehicular activities in space or for workers who require hands-free operation in hazardous environments. Our findings reveal valuable insights into the usability and efficiency of each interaction modality. Gaze, eye tracking, and voice commands demonstrated comparable performance, although their effectiveness varied depending on the individual’s eye coordination and calibration. Gestures, while intuitive, required additional refinement to ensure reliable recognition and to minimize the risk of accidental triggers. Pop-up buttons emerged as the preferred modality among participants, boasting the shortest task completion time. The implications of this work extend beyond the astronaut context. Users in various industries can benefit from effective hands-free menu navigation in both AR and VR settings. The results of this study will inform the design and implementation of future interaction systems, enabling improved efficiency and comfort for individuals operating in constrained environments.

For space missions, modeling of simultaneous physical effects (multiphysics) on the consumable resources of a spacecraft or habitat is important for early iteration on mission design and operations feasibility. These simultaneous physical effects may cause issues like toxic gas buildup, overheating, or cascading power failures that are only possible to predict using an integrated simulation of a whole spacecraft. Long duration missions may depend on this modeling for accurate prediction of usage rate of consumables and sizing requirements for spacecraft systems.In most spacecraft design efforts, high fidelity modeling of multiphysics such as simultaneous heat transfer and off-gassing is done at the component level with expensive and specialized software. Integrated simulation of the whole spacecraft is typically done by piecing together tabulated performance data from each component, which has some possibilities for missing effects from component-to-component interactions. In order to do multiphysics modeling for arbitrary spacecraft configurations in a real-time simulation with continuously varying environmental conditions, this research develops a novel simulation architecture.SimDynamX LLC and the Texas A&M ASTRO Lab have been developing Space Teams, an engineering design and simulation platform for space missions. Within Space Teams, all objects in a simulation are Entities, and they may be connected by Edges using a graph structure representation, with each Entity constituting a Node. This graph structure can change dynamically according to events that occur within the wider simulation, such as a spacecraft docking or a hatch being closed. A system of graph network algorithms are used to provide a programming framework for modeling simultaneous flow of mass and energy between Nodes. This is split into Solvers, which compute instantaneous rates from physical effects (e.g. heat conduction in Watts or fluid flow in kg/s), and Integrators which evolve the simulation forward in time, maintaining conservation laws and updating the state of the nodes. Solvers are specific to one process, such as blackbody radiation, and Integrators are specific to one type of resource quantity, such as temperature, oxygen or power.In this paper, we describe the novel architecture of Space Teams’ Advanced Resource Flow system, and compare its capabilities to other multiphysics software.

As we prepare to set foot on the Moon and subsequently turn our attention to Mars and the outer solar system, it becomes necessary to consider the entire mission environment as one unified system. One reason for this is a challenge that all space missions face: relativity. Not only will communications between a deep space vehicle and mission control be delayed, but the actual position of the Earth (and other objects) from the vehicle's point of view will begin to differ from the non-relativistic case. As it flies further and further away, this relativistic aberration will affect on-board devices that depend on precise spatial targeting, induce a drift in on-board clocks (causing them to run slower than Earth-based clocks), and even shift the apparent positions of the stars. For simulations, a particular concern is the time history of a simulated object. Under relativity, time slows down in the presence of mass (or gravity), but it also slows down for moving observers, resulting in discrepancies between observed positions, velocities, accelerations, and clocks of objects relative to each other. Overall, this implies the need for a simulation that accurately tracks these relativistic states. Presently, a relativistic simulation can be constructed for one observer using a combination of open-source tools. However, in this paper, we present an approach to simulating any number of simultaneous observers in a relativistic solar-system-scale space environment using the Space Teams simulation platform. We begin by delineating the method of simulating this type of relativistic universe, wherein each observer sees the correct positions, orientations, and relativistic aberrations of all other objects and observers, while a global, non-relativistic frame captures the overall dynamics of the system. We then discuss the use of this environment in simulating a solar-system-scale localization system, based on the Global Positioning System (GPS). We show that by tracking time histories of the spatial and temporal states of each object in the simulation, we can display an accurate view of the system from any reference point, while simultaneously applying environmental effects (such as gravitational forces) to all objects in a manner consistent with a relativistic universe.

The development of autonomous robotic platforms for space applications such as satellites, robotic manipulators and rovers has driven the need for reliable control and navigation techniques. Simultaneous Localization and Mapping (SLAM) algorithms can be employed to determine the position and orientation of robotic platforms while generating a map of an unknown hostile environment. This article presents the successful integration of ORB-SLAM2, a localization algorithm, with Voxblox, an accurate three-dimensional mapping package. This integration aims to establish a pathway for occupancy maps to be generated from feature detection algorithms. Furthermore, virtual reality is presented as an innovative solution for testing the performance of this algorithm integration in a space environment. Providing a novel simulation environment with an opportunity to diversify applications. The effectiveness of this integration and SLAM accuracy was compared to a truth model extracted from the virtual reality environment. The accuracy is dependent on the number of observed and matched features in a given frame and are intrinsic characteristic of features in the environment. The robustness of the integration is also examined through the implementation of sensor inputs in the form of stereo and RGB-D cameras. The algorithms’ integration and embedding into virtual reality allows for further developments of SLAM algorithms, to improve accuracy and robustness for autonomous navigation in any virtual environment, and extend robotic simulations and visualizations.

Space Teams Academy (STA) is a new, innovative, and virtual STEM program that teaches students about space exploration, teamwork, and engineering design principles. Recent simulation technology and advances in virtual reality (VR) make it possible to immerse and engage students in the design of advanced space exploration missions. Space Teams Academy is built upon the Space Teams platform. This sophisticated toolkit serves as a comprehensive mission design and simulation platform, providing an accurate representation of the space environment, thereby enabling users to design end-to-end virtual interplanetary missions. The Space Teams platform has been utilized for diverse space industry projects, ranging from NASA design competitions to commercial orbital servicing, and the evaluation of astronaut training scenarios. The STA program provides an unprecedented opportunity for students, from elementary through high school, to delve into an immersive and highly realistic virtual space environment. It inspires creativity, while requiring teamwork and critical thinking to solve complex problems related to the design of interplanetary vehicles, habitats on distant planets, and strategies to manage vital resources for human survival on these new worlds. The STA program is strongly aligned with the Next Generation Science Standards (NGSS) and covers a wide range of NASA's STEM Educational Objectives. Through this educational space adventure, STA students receive valuable insights from industry professionals such as space engineers, scientists and astronauts. Topics include planetary science, spacecraft design and assembly, orbital mechanics, landing on another world, and constructing extraterrestrial habitats with the ultimate goal of human sustainability. Space Teams Academy arms students with essential teamwork and critical thinking skills, while equipping them with the key STEM concepts necessary to conceptualize and execute real-world space mission scenarios. The effectiveness of STA's approach has been demonstrated by a comprehensive research study (Space Teams STEM Competition: Outcomes and Efficacy, AIAA Scitech, 2023). The study observed a significant increase in students' interest in STEM fields, a marked increase in their space knowledge and engineering skills, as well as distinct improvements in problem-solving and collaborative abilities. With the aim of furthering the impact of STA, NASA Space Grant has funded a program called Space Teams Labs (STL), which intends to reach over 10,000 students over a period of 3 years. STL seeks to provide continuous access to underserved students, promote a diverse and inclusive workforce, and to foster deeper engagement with the next generation of STEM leaders. The STL program is guided by five primary directives: strategic selection of underserved schools and programs, provision and installation of necessary hardware, fostering partnerships, training of facilitators, and most importantly, consistent education and inspiration of students. The influence of Space Teams Academy also extends beyond the younger students. The development of the program has involved hundreds of university engineering students, thus contributing to their education and skills using modern engineering tools as part of the Texas A&M Aggie Challenge program. As such, the development and implementation of Space Teams Academy is having a profound impact on students of all ages and thus shaping future careers in space exploration and other STEM fields.

The four questions on page four should read as follows below.This study aims to answer the four questions below by utilizing the quantifiable outputs described in the logic model to validate the established outcomes:1) Does the degree of engagement in Space Teams correlate with an understanding of the engineering design process?2) Does engagement in Space Teams increase interest in and positive attitudes towards space science and engineering?3) Does engagement in Space Teams increase STEM identity and self-efficacy? 4) Does engagement in Space Teams increase understanding of the importance of exploration?

Optical sensors have been used aboard satellites for a variety of purposes. In the case of proximity operations in low Earth orbit, these sensors may be used to capture images of a target satellite from a chaser. From these images state information may be obtained about the target. However, due to adverse lighting conditions caused by glare from the Sun, the images that are input to the algorithms responsible for computing state information may produce erroneous results. We propose GlareNet, a deep learning based solution to this problem. In recent years Generative Adversarial Networks (GAN) have been making strides in harnessing neural networks to generate images, and alter images to remove possible unwanted qualities, such as blur. GlareNet utilizes a conditional deep convolutional GAN (DCGAN) in order to receive an image where there is significant glare around a target satellite and subsequently remove the glare while retaining the structure of the satellite even in cases where the glare occludes parts of the target vehicle.

Current solutions for navigation and guidance on the lunar South pole have limited visual spatial cues to help guide astronaut crewmembers to mission objectives. Challenges include low solar elevation angles that create areas of high contrast, permanently shadowed regions (PSR), and a lack of radionavigation systems such as the Global Positioning System (GPS) on Earth. This paper proposes a novel and adaptive navigation system through the Space Communications, Operations, and User Telepresence (SCOUT) Augmented Reality (AR) assistant to support astronaut navigation on the Moon. SCOUT is an artificial intelligence (AI) system that can be integrated into an astronaut's visor to collaboratively identify points of interest, detect hazards, and guide a group of crewmembers towards mission waypoints using hologram overlays. SCOUT uses deep learning to iteratively improve its under-standing of the nearby environment, and can be trained using data from robotic systems deployed on the Moon before astro-nauts arrive. To assist an astronaut's navigation capability in extreme environments, SCOUT displays 3D spline holograms towards waypoints and objectives. For example, SCOUT can display a clear label during an extra-vehicular activity (EVA), such as a switch that should be flipped, or which direction they should walk. To effectively navigate through and explore any region of the lunar surface, SCOUT has been trained on a multitude of randomly generated Moon-like virtual environ-ments. Through this process, this AI system can quickly identify features that it has seen before as well as adapt to new situations it has never experienced before. Results show that SCOUT can map out an optimal path towards a waypoint, even without having previous data samples from that specific environment. SCOUT is designed to benefit the user as much as possible, through automated guidance and continuous learning of the surrounding environment.

The recent advancement of government and commercial spaceflight programs have endowed the public with a re-newed excitement for the colonization of Mars. This paper proposes a concept for a sustainable Mars habitat architecture, the Joint Enhanced Architecture for Navigating New Environments (JEANNE). This architecture focuses on safety of the human crew through remotely operated construction, countermeasures against physical & psychological risks, and the Computerized Relationship Layout Planning (CORELAP) methodology for habitat layout. The main challenges considered include: radiation exposure, remote habitat construction, power generation, waste management, human physiology in 3/8's of Earth's gravity, and the behavioral and psychological performance of the crew on Mars, to name a few. The JEANNE Habitat aims to address the obstructions to human life on Mars using a hybrid of low- and high-Technology Readiness Level (TRL) technology, the Scalable Interactive Model of an Off-World Community (SIMOC), CORELAP, In-Situ-Resource-Utilization (ISRU), and overarching NASA standards and requirements. The scenario described in this paper is focused on a 500 to 600 day Mars surface mission in the Arcadia Planetia region with a four-person crew. Before the crew launches in 2033, there is a planned cargo launch in 2031. 3D models for each habitat module were designed and plugged into the SIMOC simulator. This C++ based simulation includes verification and testing of the design against the requirements that a user specifies. The primary points of interest in the simulation are on crew safety, ISRU, atmospheric management, water management, food production, and waste management. Results from the SIMOC simulation prove that the proposed design can support a crew of four inhabitants for the duration of the mission. The habitat status after 600 days yields the internal atmospheric composition of 19.2% oxygen and 0.05% carbon dioxide. For contingency scenarios, 1420 kg of potable water and 585 kg of food remain. The CORELAP algorithm defined several layout requirements, namely the bathroom must be kept far away from the food areas to prevent contamination while also being kept near the greenhouse. Also, the exercise room must be far away from work and sleeping areas due increased noise levels. Through design, analysis, and simulation, the JEANNE habitat is set to be ready for 2033 human surface missions, meeting NASA's Journey to Mars goal.

Long duration spaceflight missions will require novel exercise systems to protect astronaut crew from the detrimental effects of microgravity exposure. The SPRINT protocol is a novel and promising exercise prescription that combines aerobic and resistive training using a flywheel device, and it was successfully employed in a 70-day bed-rest study as well as onboard the International Space Station. Our team created a VR simulation to further augment the SPRINT protocol when using a flywheel ergometer training device (the Multi-Mode Exercise Device or M-MED). The simulation aspired to maximal realism in a virtual river setting while providing real-time biometric feedback on heart rate performance to subjects. In this pilot study, five healthy, male, physically-active subjects aged 35 ± 9.0 years old underwent 2 weeks of SPRINT protocol, either with or without the VR simulation. After a 1-month washout period, subjects returned for a subsequent 2 weeks in the opposite VR condition. We measured physiological and cognitive variables of stress, performance, and well-being. While physiological effects did not suggest much difference with the VR condition over 2 weeks, metrics of motivation, affect, and mood restoration showed detectable differences, or trended toward more positive outcomes than exercise without VR. These results provide evidence that a well-designed VR “exergaming” simulation with biometric feedback could be a beneficial addition to exercise prescriptions, especially if users are exposed to isolation and confinement.

As self-sustainability gains importance for future human spaceflight missions, in-situ resource utilization (ISRU) precursor missions can provide a low-risk option to ensure availability of resources such as water and oxygen. This paper proposes a design to create a fully autonomous, modular robotic architecture for the refining and processing of icy regolith into water in the Permanently Shadowed Regions of the Lunar South Pole. The contributions made in this paper include an in-depth design and analysis of a low-risk lunar south pole ISRU mission architecture. Hardware included in this design are the three types of rovers: excavation, collection, and transportation. All rovers share a common main body, which contains hardware systems for the drivetrain, position and attitude determination, communication, imaging, thermal management, power delivery, and swappable batteries. These rovers operate with mesh-network communications and use swarm logic to aid in path planning and collision avoidance. This design was verified through simulations of a typical hour-long “work cycle” that includes all the steps of the water excavation, refining, and delivery. Per-work-cycle performance metrics were extrapolated for the mission duration to calculate total mission performance.

As the scope of human spaceflight continues to expand, the Human Systems Integration (HSI) developed to support complex missions must be robust and efficient. One of the most critical elements of any human spaceflight mission is training flight operations teams such that they can carry out their objectives effectively. As more distant destinations such as the Moon or Mars are targeted for human spaceflight, ensuring crew have the tools they need to overcome new types of challenges will be a significant focus of the training infrastructure. A method being developed to support this infrastructure is the Simulation Builder, Analysis, and Development (SimBAD) Tool. SimBAD uses the SpaceCRAFT simulation platform to construct high-fidelity simulations and develop procedures with which crew can use to train for missions in Virtual Reality (VR). Using this tool, two simulations are built as conceptual demos for how SimBAD and tools like it will be used on future missions to address knowledge gaps associated with risk of inadequate HSI.

The reality of space exploration is that it is difficult and expensive. In recent years, increasingly realistic simulation tools paired with virtual reality have made space system design and mission planning much more accessible to people outside of leading aerospace companies by removing the need for costly and highly specialized physical equipment. These advancements have made it possible to begin teaching aerospace topics to children in elementary, middle, and high school, as well as introduce them to proper engineering techniques in the interest of bolstering their STEM education. In order to engage the students, a competition geared toward teaching engineering concepts can be used to bring teams of students together to work on a simplified aerospace problem. This paper discusses possible implementations of such a competition using a high fidelity real-time engineering simulation program.

This paper proposes a new method for robotic teleoperation scenarios with significant time delay utilizing Extended Reality (XR) technologies, specifically for Lunar base assembly and maintenance. Traditional teleoperation interfaces interpret commands sent through a desktop computer screen and are limited in providing spatial awareness cues. Additionally, the software often has difficulty in accounting for a time delay or providing intuitive techniques for human-robot interaction (HRI). This paper presents a new XR interface for robotic teleoperation using a predictive 3D simulated virtual environment developed using the SpaceCRAFT platform. SpaceCRAFT is a multiuser Virtual Reality (VR) and Augmented Reality (AR) systems engineering toolbox software that allows for worldwide collaboration on design and test of space systems. SpaceCRAFT provides the operator a virtual interface to the robot represented through three different states, a Virtual Command State (VCS), Predicted Current State (PCS), and Time Delayed State (TDS). To improve situational awareness and the PCS representation of the robot, SpaceCRAFT processes LIDAR and RGB camera data to create a constantly updating 3D point cloud representation of the remote environment. An experiment involving teleoperation of a 3D printed robotic manipulator was run, with a simulated time delay similar to that between the Earth and the Moon. Results demonstrate that the predictive approach provides an effective and flexible method of robot teleoperation that helps achieve successful task completion, even with a large time delay.

During extravehicular activities (EVAs), astronauts are heavily dependent on the Mission Center (MCC) and their Intra-Vehicular Astronaut (IVA) counterparts. Each procedure step in a mission is relayed to the astronaut through a real-time voice loop, and emergency procedures are written on cuff checklists that astronauts must read from their spherically shaped helmet. In all situations, crew members heavily rely on IVA or MCC support, especially when they do not understand a procedure or need help with a specific problem. However, it can be hard to communicate procedures effectively due to a lack of visual diagrams and situational awareness between the two parties. To improve EVA efficiency, we investigated the use of a virtual whiteboard on a heads-up display during a lunar surface EVA task with virtual reality (VR). The virtual whiteboard allows MCC to send additional visual guidance (e.g., drawings and annotations) overlayed on the astronaut’s visual field of view to better assist with mission tasks. We conducted a between-subjects experiment where 21 participants were asked to accomplish a rover repair procedure, with ( n = 11) and without ( n = 10) the virtual whiteboard, using a VR lunar environment with support of a research proctor acting as MCC. The whiteboard group completed the rover procedure 39.1% faster than the non-whiteboard group, and this difference was statistically significant ( p = 0.017). The total number of words exchanged during the experimental sessions was not statistically different between groups ( p = 0.99). However, participants in the whiteboard group showed a tendency to talk less than their counterparts, while the research proctor in the whiteboard group showed a tendency to speak more. Finally, analysis of the spatial locations during the experiment indicated that whiteboard participants stayed closer to the rover, showing a better focus on the task at hand and therefore short completion times. The results of this experiment inform future development of AR spacesuit technologies for future planetary exploration EVA operations.

The exploration of space will require ever-increasing exposure to microgravity environments. The human response to this exposure has been categorized and mitigated via countermeasures, principally exercise. However, additional constraints to future mission design minimizes the allotted space and modalities for exercise, creating a risk for psychological fatigue, a reduction in motivation, and a suite of other categorical factors that could, taken together, present a risk for reduced adherence to the countermeasures and/or mission performance. Thus, the current study will examine the effects of a virtual reality (VR) intervention on spaceflight-validated exercise protocols using a prototype rowing ergometer designed to operate within the constraints of future long-duration exploration missions (LDEM). The Integrated Resistance and Aerobic Training (dubbed "SPRINT") protocol will be used in conjunction with a combination flywheel and resistance training device (M-MED) utilized in prior bedrest studies. The SPRINT protocol trades exercise duration for intensity, providing similar benefits to existing countermeasures while reducing time spent on exercise. The M-MED permits resistance training on the muscles most effected by microgravity exposure on the same device used to train cardiovascular function, thus reducing the volume and weight requirements of the exercise countermeasure. It is upon this framework that we will add the VR rowing simulation. VR has shown to be a lightweight, reliable, and enjoyable technology in numerous studies, while exergaming has been shown to improve measures of motivation and adherence. We will create a rowing simulation that can integrate with a rowing ergometer and any exercise protocol, and then implement it on the M-MED with SPRINT. The simulation will feature virtual teammates, virtual competitors, and other gaming mechanisms that encourage a user to maintain a prescribed heart rate intensity in a way that aims to maximize factors associated with enjoyment and adherence. We plan to conduct a within-subjects experiment on an astronaut-like population. Subjects will be randomly assigned to VR or non-VR in their initial experiment, complete the SPRINT protocol on the M-MED, break for a one-month minimum washout period, then return to complete the protocol again in the other group. As a pilot study, dependent variables have been selected broadly. Physical and psychological outcomes are to be measured alongside adherence to and motivation toward this very challenging protocol. Additional measures are to be made of virtual presence, preexisting bias toward or against VR, and personality traits, which may influence a preference for or against VR. Preliminary data on non-VR subjects shows increasing measures of state-trait anxiety, negative feelings toward the exercise, and amotivation from the start to the end of the protocol. It is hypothesized that overall attitudes toward the protocol will improve with the VR intervention as indicated by metrics of adherence, motivation, affect, and mood restoration. The results of this study will inform future designs of exercise combined with VR applications for implementation during LDEM.