Distributed Collaboration and Coordination for Planetary Exploration Mission Support

When humans next walk on the surface of the Moon or Mars, they will be supported by a vast infrastructure of experts on the ground in Mission Control Center (MCC) and elsewhere, robotic assets with them on the surface, and computer systems analyzing data. The act of donning a spacesuit and stepping out onto the surface is inherently risky, and as such the entire system of extravehicular activity (EVA) decision makers will be focused on ensuring the crew is safe and productive while they explore. The fundamental problem this system will face is that of uncertain communications impinging on the ability of the decision makers to share knowledge, plans, and goals in real-time. This research proposes to systematically address uncertain communication by developing a set of techniques and technologies for distributed collaboration and coordination that will be effective even when astronauts and their support systems cannot communicate. The research objectives will be accomplished by building off the current state of the art in autonomy and AI. There has been a well-documented history of autonomous robots successfully exploring extreme environments such as the deep ocean. These exploration campaigns sometimes include multiple vehicles using centralized collaboration algorithms to cooperate about goals and plans while ensuring safe execution. Other autonomous planners and schedulers have demonstrated the ability to generate new plans on-the fly and proactively prepare for contingencies. Most recently there has been work on the temporal network algorithms that adds the ability to model uncertain communications and still reason over the feasibility of plans. This research primarily proposes to integrate the state of the art of uncertain communications in temporal constraint algorithms with existing capabilities for multi-agent task and motion planning with bounded-risk. The contribution will be a set of algorithms for distributed collaboration and coordination under uncertain communication. Secondarily, reasoning about the intent of other agents when communication is limited will enable agents to probabilistically model the actions their peers will take. By using studies of how astronauts behave on EVA and how field geologists behave in the field, this research will contribute notional intent-recognition for contingency generation and online planning for EVAs. Combined, the contributions of this research will enable astronauts and their support teams to safely and effectively operate even when they cannot talk to one another. Distributed reasoning over uncertain communications is a fundamentally new capability that, while being framed in terms of utility for human spaceflight exploration, is applicable across a wide variety of NASA interests where remote agents must make decisions. Notably, deep-space autonomous robotic exploration is largely dependent on the same body of planning and scheduling algorithms to which this research proposes to contribute. It is a low TRL technology that could be applied to projects such as CAVES at JPL, which is developing an autonomous agent that is capable of operating in lava tubes. Likewise, this research could ultimately enable more robust communication for lunar rovers like VIPER that will be exploring permanently shadowed regions on the Moon.

Distributed reasoning over uncertain communications is a fundamentally new capability that, while being framed in terms of utility for human spaceflight exploration, is applicable across a wide variety of NASA interests where remote agents must make decisions.

This is early/mid-stage (TRL 3) — the most realistic path in is NASA SBIR/STTR, which funds small businesses and research institutions to develop technology aligned with NASA's needs (equity-free, phased funding). Check whether a current SBIR/STTR solicitation topic overlaps with this project's technology area, or contact the project directly (above) to ask.

None of these are guaranteed paths for this specific project — TechPort itself doesn't have an "apply" button. Reaching out to the contact(s) above with a specific question is usually the fastest way to find out what's actually open.