🔭 astrophysics

Planetary Exploration 3.0: A Roadmap for Software-Defined, Radically Adaptive Space Systems

This paper proposes "Planetary Exploration 3.0," a new paradigm utilizing software-defined, radically adaptive space systems with onboard intelligence to enable single missions to autonomously evolve and conduct both exploratory and hypothesis-driven science in the distant, unexplored outer Solar System.

Original authors: Masahiro Ono, Daniel Selva, Morgan L. Cable, Marie Ethvignot, Margaret Hansen, Andreas M. Hein, Elena-Sorina Lupu, Zachary Manchester, David Murrow, Chad Pozarycki, Pascal Spino, Amanda Stockton, Math

Published 2026-04-24

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Masahiro Ono, Daniel Selva, Morgan L. Cable, Marie Ethvignot, Margaret Hansen, Andreas M. Hein, Elena-Sorina Lupu, Zachary Manchester, David Murrow, Chad Pozarycki, Pascal Spino, Amanda Stockton, Mathieu Choukroun, Soon-Jo Chung, John Day, Alexander Demagall, Anthony Freeman, Chloe Gentgen, Michel D. Ingham, Charity M. Phillips-Lander, Richard Rieber, Alejandro Salado, Maria Sakovsky, Lori R. Shiraishi, Yisong Yue, Kris Zacny

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are planning a trip to a completely unknown island.

The Old Way (Planetary Exploration 2.0): The "Specialized Panda"
Currently, NASA explores space like a highly specialized panda. If you want to see a bamboo forest, you send a panda. If you want to see a bamboo forest again but with a better camera, you send a slightly smarter panda. If you want to eat the bamboo, you send a panda with a special mouth.
This worked great for Mars. We sent a flyby, then an orbiter, then a lander, then a simple rover, and finally a super-complex rover. We learned a little bit, then sent a new mission to learn a little more. It's like a slow, careful staircase.

But this doesn't work for the deep outer solar system (like Neptune or the Oort Cloud). The trip takes 10 years. If you send a "scout" and it comes back with a surprise, you can't just send a "follow-up" mission 10 years later. By then, you've lost a generation of scientists, and the mission is too expensive. You need to get it right the first time.

The New Way (Planetary Exploration 3.0): The "Adaptable Rat"
The authors propose a new paradigm called Planetary Exploration 3.0 (PE 3.0). Instead of a specialized panda, they want to send a rat.
Rats are generalists. They can eat almost anything, live in sewers or attics, and adapt to whatever environment they find. They don't need a specific plan; they just react to what's in front of them.

In PE 3.0, we send one single mission (or maybe two) that is designed to change its own mind and its own body while it's flying.

How Does a Spacecraft "Change Its Mind"?

The secret sauce is Software-Defined Space Systems.

Think of a traditional spacecraft like a Swiss Army Knife that is glued shut. You buy it, and it has a knife, a screwdriver, and a toothpick. If you need a hammer, you're out of luck. The hardware is fixed.

A PE 3.0 spacecraft is like a 3D-printing robot made of liquid metal.

Hardware: Instead of fixed antennas, it has "smart" antennas that can change shape with a software update. Instead of a fixed camera, it has a lens that can zoom, focus, or change colors instantly.
Software: The brain of the spacecraft isn't just following a script. It's an AI scientist. If it lands on a planet and finds a rock that looks weird, it doesn't wait 4 hours for a signal from Earth to say, "Hey, look at this rock!" It decides on its own, "This rock is interesting! I'm going to drill into it and analyze the gas inside."

The Three Big Changes

The paper outlines three major shifts in how we think about space travel:

From "Checklist" to "Backpacking Trip":
- Old Way: A business trip. You have a strict schedule: 9:00 AM meeting, 10:00 AM lunch. You pack exactly what you need.
- New Way: A backpacking trip. You don't know what you'll find. So, you pack a Swiss Army knife, duct tape, and rope. You don't know exactly what you'll use them for, but you know they'll come in handy. The spacecraft is packed with "general purpose" tools that can be repurposed on the fly.
From "Hardware Rules" to "Software Rules":
- Old Way: The rules are written in concrete. If the antenna breaks, the mission is over.
- New Way: The rules are written in code. If the antenna breaks, the software says, "Okay, I'll use the backup radio to send data, and I'll use the camera to help navigate." The spacecraft can rewire itself virtually.
From "Perfect Plan" to "Learning as You Go":
- Old Way: You plan every step before you leave Earth.
- New Way: You have a goal (e.g., "Find life"), but you don't know the steps. The spacecraft explores, learns, forms a new hypothesis, and then tests it—all while flying.

What Would This Look Like in Real Life?

The paper suggests three cool missions that could only happen with this new tech:

The Neptune/Triton Flyby: Imagine flying past Neptune's moon, Triton. You see a giant geyser shooting water into space. In the old days, you'd just take a picture and fly away. In PE 3.0, the spacecraft says, "Whoa, that's a plume! I'm going to change my course, dive into the plume, and taste the water to see if it has life in it." It does this in seconds, without waiting for Earth.
The Ocean World Explorer: Imagine a robot that lands on an icy moon (like Enceladus). It melts through the ice, swims in the ocean, and then climbs back out. If it gets stuck in the mud, it doesn't call for help. It figures out a new way to wiggle free using its "snake-like" body. It might even build a small bridge out of ice to get to a better spot.
The Oort Cloud Scout: This is the edge of our solar system, where comets live. We know almost nothing about it. A PE 3.0 mission would be a "mother ship" that drops off little "daughter" robots. If the mother ship sees a weird comet, it tells the daughter robots, "Go check that one out!" If a robot breaks, the others reconfigure themselves to do the job.

The Bottom Line

We are moving from an era of careful, step-by-step exploration (which works for Mars) to an era of bold, one-shot, shape-shifting exploration (which is necessary for the deep dark of space).

We are trading the safety of a rigid plan for the power of a smart, adaptable machine that can think, learn, and change its own body to survive the unknown. It's the difference between sending a soldier with a fixed map and sending a master explorer with a compass, a map that redraws itself, and the ability to build a bridge if the road is gone.

Based on the provided paper, here is a detailed technical summary of "Planetary Exploration 3.0: A Roadmap for Software-Defined, Radically Adaptive Space Systems."

1. Problem Statement

Current planetary exploration relies on Planetary Exploration 2.0 (PE 2.0), a paradigm characterized by incremental, multi-mission campaigns (e.g., NASA's Mars exploration). This approach involves a sequence of missions: low-risk flybys $\rightarrow$ orbiters $\rightarrow$ landers $\rightarrow$ rovers, where each mission builds on data from the previous one to refine hypotheses and requirements.

PE 2.0 is untenable for the outer Solar System and beyond due to:

Prohibitive Cruise Times: Missions to outer planets (e.g., Saturn, Neptune) or the Oort Cloud take decades. An incremental approach would require centuries to achieve deep exploration.
Deep Uncertainty: Unlike Mars, outer Solar System targets (icy moons, Kuiper Belt Objects, Oort Cloud) have little to no prior proximity data. Critical environmental factors (terrain mechanics, plume dynamics, surface composition) are unknown, making pre-defined requirements and static mission architectures risky and often infeasible.
Communication Latency: Deep space missions suffer from long communication delays and intermittent connectivity, rendering ground-in-the-loop adaptation too slow for real-time decision-making during critical events (e.g., plume flybys or landing).

2. Methodology and Proposed Paradigm

The paper proposes Planetary Exploration 3.0 (PE 3.0), a paradigm shift where a single mission (or a very small number of missions) explores unvisited worlds using Radically Adaptive Space Systems. Instead of spreading adaptation across multiple missions over decades, PE 3.0 embeds adaptability into a single spacecraft that evolves its capabilities in situ.

The methodology is built on three core paradigm shifts:

From Hypothesis-Driven to Capability-Driven Missions: Moving away from static Science Traceability Matrices (STMs) that link fixed hypotheses to specific instruments. PE 3.0 missions use "Swiss Army knife" instruments designed for broad utility, allowing the spacecraft to dynamically form and test hypotheses based on real-time data.
From Hardware-Defined to Software-Defined Systems (SDSS): Transitioning from fixed hardware architectures to systems where functions are defined and reconfigured via software. This includes reconfigurable hardware, multi-functionality, and modularity.
From Requirement-Based to Affordance-Based Engineering: Shifting from rigid, pre-defined requirements to design principles based on "affordances" (what a component can do in unforeseen contexts) and set-based design, allowing for deferred commitment and continuous redesign.

3. Key Contributions

A. Systems Engineering Framework (Section II)

Dynamic Science Definition: Proposes replacing static STMs with conditional plans and decision trees that update as data is acquired. It advocates for "AI scientists" capable of autonomous hypothesis generation and data prioritization.
Adaptive Requirements: Introduces the concept of System Adaptivity Level (SAL), a metric analogous to Technology Readiness Level (TRL) but measuring a system's ability to tolerate uncertainty and repurpose functions.
Design Principles: Derives eight key principles for adaptive missions, including:
- Extra Sensing Modes: Redundant sensors for anomaly diagnosis.
- Redundant Degrees of Freedom (DOFs): Mechanical flexibility for alternative locomotion or manipulation.
- Interconnectivity: Flexible data/power routing to allow subsystem repurposing.
- Resource Margins: Excess fuel, power, and compute to enable adaptation.
- Onboard Reprogrammability: The ability to update flight software and reconfigure hardware in flight.
Verification & Validation (V&V): Shifts focus from verifying compliance with fixed requirements to validating capabilities across a wide range of scenarios using adversarial testing and AI-assisted simulation.

B. Enabling Technologies (Section III)

The paper surveys technologies required to build Software-Defined Space Systems:

Reconfigurable Hardware: Software-defined radios (SDR), active phased array antennas (APA) for dynamic beamforming, adaptable thrusters (e.g., Hall effect thrusters with tunable specific impulse), and morphing thermal radiators.
Multi-functionality: Repurposing structural elements for thermal protection or energy storage (structural batteries) and using navigation cameras for science.
Software-Defined Robotics: High-DOF robots (e.g., snake-like robots) that can switch gaits (rolling, slithering) to escape traps, and modular/swarm robotics that can self-assemble or reconfigure.
In-Space Manufacturing: Additive manufacturing and modular assembly to allow spacecraft to grow, repair, or reconfigure physically in space, potentially utilizing in-situ resources (ISRU).

C. Onboard Intelligence (Section IV)

Autonomous Science: AI for real-time data triage, anomaly detection (e.g., identifying fresh impacts or plumes), and autonomous experiment planning.
Embodied AI: Control systems that learn from physical interaction (visual, proprioceptive, and tactile feedback) to adapt mobility and manipulation strategies in unknown terrains.
Self-Verification: AI capable of verifying its own knowledge and confidence levels before transmitting data to Earth.

D. Mission Concepts (Section V)

The paper illustrates PE 3.0 through three concept studies:

Neptune/Triton Smart Flyby: A mission using autonomous navigation to target cryovolcanic plumes on Triton. The spacecraft dynamically adjusts trajectory and instrument settings to sample plume composition, a task impossible with a fixed pre-planned trajectory.
Ocean World Explorer: A mission to an icy moon (e.g., Enceladus) involving an orbiter, a lander, and a subsurface probe. It utilizes modular, reconfigurable robots (e.g., screw-finned bots) to transition from surface mobility to ice melting and subsurface swimming, adapting sample collection strategies based on real-time vent conditions.
ORIGIN (Oort Cloud Reconnaissance): A "mother-daughter" architecture where a mothership performs broad reconnaissance and deploys specialized daughtercraft. The system autonomously balances exploration (searching for objects) and exploitation (targeting specific bodies for detailed study) in an environment with almost zero prior knowledge.

4. Results and Findings

Feasibility of One-Shot Exploration: The study concludes that PE 3.0 is not just an option but a prerequisite for exploring the outer Solar System, as the time scales of incremental exploration are incompatible with human lifespans and mission funding cycles.
Technology Readiness: While many enabling technologies (SDR, FPGAs, autonomous navigation) exist at moderate TRLs, significant gaps remain in high-DOF robotics, in-space self-manufacturing, and robust embodied AI for unstructured environments.
Systems Engineering Shift: The paper demonstrates that traditional systems engineering (rigid requirements, fixed STMs) must be fundamentally restructured to support continuous, in-flight adaptation. The proposed "Affordance-Based" and "Set-Based" design methods provide a theoretical framework for this transition.

5. Significance

Paradigm Shift: PE 3.0 represents a fundamental change in how space exploration is conceived, moving from a "predict and verify" model to a "sense, adapt, and learn" model.
Scientific Return: By enabling a single mission to perform the roles of multiple PE 2.0 missions (exploration, characterization, and hypothesis testing), PE 3.0 maximizes scientific return per dollar and time, particularly for high-risk, high-reward targets like ocean worlds and interstellar objects.
Resilience: Adaptive systems are inherently more resilient to anomalies and environmental surprises, reducing the risk of total mission failure due to unexpected conditions (e.g., rocky terrain or brittle rock).
Future Roadmap: The paper serves as a roadmap for NASA and the global space community, identifying specific technology development areas (autonomy, reconfigurable hardware, modular robotics) required to realize the next era of deep space exploration.

In summary, the paper argues that to explore the unknown frontiers of the Solar System, humanity must stop building "pandas" (highly specialized, fragile systems) and start building "rats" (generalist, adaptable, software-defined systems) capable of surviving and thriving in unpredictable environments.