Real mission narratives, astronaut profiles, step-by-step scenario walkthroughs, and STEM learning modules — everything you need to understand how ALRS saves lives on the Moon.
Detailed records of every ALRS operational deployment — from the first alert to final extraction.
Mission day 12 began as a routine geological survey of the Taurus-Littrow Valley's basaltic plains. At T+06:22, Cian McMorna's suit telemetry broadcast an anomaly: a micro-fracture in the secondary CO₂ scrubber had propagated under thermal cycling, silently spiking CO₂ partial pressure from 0.3 kPa to a life-threatening 5.1 kPa. With Earth comm-lag at 1.3 seconds each way, every second mattered.
Conducting a water-ice core sampling inside Shackleton's permanently shadowed region, Dr. Maya Chen slipped on a steep regolith slope and slid 40 metres, disabling her primary comm antenna on impact. The ALRS relay mesh detected the communication gap at T+00:09 and immediately activated the Dead-Reckoning Rescue Protocol — operating entirely without Earth confirmation.
A high-energy solar proton event corrupted Yuki Tanaka's rover navigation firmware, leaving the vehicle stranded 18 km from Artemis Base. The radiation burst also knocked out GPS for a 6-hour window. ALRS pivoted to star-tracker celestial navigation, a backup mode never previously used in an active rescue scenario.
The remarkable individuals who trusted ALRS with their lives — and the data that keeps them safe.
"Every rock I lift here is a chapter in the Solar System's autobiography. I'm just the librarian."
"The ice in Shackleton may hold clues to life elsewhere. Danger is simply the admission price."
"I design systems to save others. Being saved by one is humbling — and deeply educational."
Step-by-step breakdowns of ALRS autonomous decision chains — every action, every probability shift.
A micro-fracture in the CO₂ scrubber triggered a cascade failure across secondary life-support subsystems. This walkthrough documents every autonomous decision made by ALRS over 107 minutes, the probabilistic reasoning behind each action, and the lessons encoded into the system thereafter.
When Dr. Chen's primary comm unit failed inside the crater's permanently shadowed region, ALRS faced its hardest constraint: zero telemetry from the astronaut, zero GPS, zero Earth confirmation. This is the story of how the Dead-Reckoning Rescue Protocol earned its first operational validation.
Deep dives into the science, engineering, and economics behind autonomous lunar rescue.
ALRS uses a modified A* algorithm operating on a 1-metre-resolution digital elevation model derived from LRO LOLA data. Each node in the graph carries a traversal cost weighted by slope angle, surface roughness (derived from shadow analysis), and proximity to known hazard zones. The heuristic function is a weighted Euclidean distance to the goal, adjusted for predicted soil shear strength.
Stereo cameras on rescue vehicles produce dense point clouds at 10 Hz. A GPU-accelerated plane-fitting algorithm segments traversable ground from obstacles in under 80 ms per frame. New obstacle data updates the navigation graph dynamically, allowing the rover to re-plan mid-traverse without stopping.
When two or more vehicles operate simultaneously, ALRS uses a decentralised auction protocol: each vehicle bids for waypoints based on its current battery state, distance, and payload capacity. The highest-utility assignment is made globally, preventing redundant coverage and resource conflicts.
Seven biosensors (heart rate, SpO₂, skin temperature, CO₂ pCO₂, suit pressure, hydration index, and galvanic skin response) feed a Long Short-Term Memory (LSTM) neural network trained on 14,000 hours of simulated EVA data. The model outputs a "risk score" every 500 ms; a score above 0.72 triggers automatic alert escalation.
Rather than reacting to failures, ALRS forecasts them. A gradient-boosted regression model predicts O₂ consumption rate, CO₂ scrubber saturation time, and suit thermal load 30 minutes ahead. This early-warning capability allows preventive resource deployment — often before the astronaut notices any discomfort.
Each astronaut's model is fine-tuned during a 72-hour ground-truth calibration period before their mission. Transfer learning from the global model is applied, then personalised with individual physiological signatures. This reduces false-positive alert rates by 61% compared to generic threshold systems.
At 1.62 m/s² (16.5% of Earth gravity), an astronaut in a 130 kg EVA suit has an effective surface weight of ~21 kg. This allows greater agility but reduces traction — a crucial parameter for rover navigation algorithms. Slope traversal limits differ significantly from terrestrial models: ALRS uses a modified Bekker-Wong soil mechanics model calibrated for anorthositic regolith.
Lunar regolith consists of angular, glassy particles produced by 4.5 billion years of micrometeorite bombardment. Unlike rounded terrestrial soils, regolith has very low cohesion but high internal friction. ALRS terrain models account for the bimodal particle size distribution and electrostatic levitation effects that can cause fine particles to cling to optical sensors and solar panels.
EVA suits operate at 101.3 kPa internal pressure (pure O₂ at ~29.4 kPa in some configurations). A micrometeorite puncture as small as 0.5 mm can cause pressure loss of ~0.5 kPa/min. ALRS models pressure decay using the Hagen-Poiseuille leakage formula and dispatches repair resources calibrated to the breach size inferred from sensor data.
NASA's safety regulations apply a Value of Statistical Life (VSL) of $11.6 million per astronaut, derived from federal regulatory standards. With training costs exceeding $50M per mission-ready astronaut and decades of institutional knowledge embedded in each crew member, the economic case for robust autonomous rescue systems is overwhelming — even before humanitarian considerations.
A rescue that succeeds within 2 hours preserves 80–90% of mission objectives. A 24-hour delay caused by inadequate response infrastructure can cascade into $200–400M of lost science, hardware, and re-scheduling costs. ALRS's <2-minute autonomous response time directly protects mission investment downstream.
Amortised over a 10-year operational period with 8 missions per year, ALRS costs approximately $1.3M per rescue capability activation. When compared to the alternative — a failed rescue leading to astronaut loss and mission abort — the net present value of the system exceeds $340M in risk-adjusted terms. Independent NASA actuarial models confirm a 340% ROI over the programme lifecycle.
A chronological record of the ALRS programme from concept approval through active operations.