L2-CDH-AUTO CRITICAL COMMAND CONTROL Level 2 · software

Autonomy and Decision Engine

Autonomous Decision Engine (ADE)

The Autonomous Decision Engine (ADE) is an onboard computational subsystem within the Lunar Ark command and data handling architecture that executes mission planning, machine learning inference, and anomaly response. The ADE uses hierarchical task planning, model-based reasoning, and reinforcement learning to manage Ark operations and allocate resources without continuous ground intervention. Fully autonomous control is necessitated by the 1.3-second one-way communication latency to Earth, which prevents remote intervention during time-critical anomalies. Operating across a 100-year mission lifetime requires the system to adapt to novel situations and autonomously manage resources through recurring lunar day and night cycles while balancing proactive intervention with conservative safe-mode protocols for preservation-critical systems.

AI/ML inference, mission planning, anomaly response, and goal-directed autonomous behavior

Purpose

Make high-level autonomous decisions about Ark operations and resource allocation

Context

Component of L1-CDH within the Lunar Ark system

Principles

▸Goal-directed autonomy enables multi-year independent operation without ground-in-the-loop
▸Model-based reasoning uses system models to predict failures and plan responses
▸Hierarchical task planning decomposes high-level goals into executable action sequences
▸Reinforcement learning enables adaptation to novel situations over mission lifetime
▸Safe-mode hierarchy ensures graceful degradation when autonomy confidence is low

Typical implementations

▸NASA IDEA (Integrated Distributed Executive Architecture) for autonomous spacecraft
▸JPL AEGIS autonomous science targeting on Mars rovers (Curiosity, Perseverance)
▸ESA PROBA satellite constellation autonomous operations
▸Timeliner automated procedure execution on ISS

Lunar considerations

▸Communication delays (1.3s one-way) make full autonomy essential for time-critical decisions
▸100-year mission requires learning and adaptation beyond initial programming
▸Must balance autonomous action with conservative safe-mode for preservation-critical systems
▸Lunar day/night cycling creates recurring autonomous resource management challenges

Specifications

Functional

primary function

Make high-level autonomous decisions about Ark operations and resource allocation

Operational

thermal range c	-173, 127
lifetime years	100

Interfaces

Provides

→Central Processing Systemdatacritical
High-level action commands and task schedules

Requires

←Sensor Fusion Enginedatacritical
Fused sensor state for decision-making
←Flight Software Frameworkdatacritical
Flight software framework and API interfaces