{}

• Modern rideshare missions deploy dozens to hundreds of CubeSats in a single separation sequence.
• Common launchers: Falcon 9 Transporter (SSO), PSLV, Vega, Soyuz, Electron; altitude typically 450–600 km SSO.
• Aggregators: Exolaunch (EXOpod), Planetary Systems (CSD), Spaceflight Inc., ISILAUNCH; deployers ensure mechanical and electrical isolation.

• PPOD/QuadPack, CSD, EXOpod, and similar rail/deployer interfaces per Cal Poly CubeSat standard.
• Requirements: no protrusions, kill-switches engaged, remove-before-flight (RBF) pins until integration.
• Spring ejection induces tumble; satellites must detumble autonomously (e.g., B-dot control) within minutes.

• Licensing: FCC/NTIA (US) or national regulator; ITU filings by sponsoring administration.
• Amateur bands require IARU coordination; unique call signs and beacon IDs for identification in crowded deployments.
• Debris mitigation: deorbit compliance (≤25 years traditional, trend toward ≤5 years); drag augmentation recommended.

• Radio silence period often mandated (e.g., ≥30 min) to avoid interference during separation; then low-duty beacon.
• Initial acquisition uses generic TLE clouds; identification via unique beacon patterns and doppler signatures.
• ADCS detumble, transition to Safe by default; power-up rails and payloads only after health checks.

• Minimal ground contact initially; autonomy must manage power, thermal, and comms without commands.
• Prioritized downlink: beacon frames include health summary, fault events, and basic ephemeris.
• Conjunction risk increases; timely ephemeris updates and identifiable beacons improve cataloging.

• Default Safe mode on boot; detumble enabled; transmit inhibited until post-separation timer expires.
• Beacon content: subsystem vitals, fault counts, mode, unique ID; short frame size to fit low bitrate.
• Battery charge rules enforced (0–45 °C) from first boot; payloads disabled until margins verified.
• Hysteresis prevents oscillation in crowded RF; rate limits on restarts and mode changes.

• Day/night cycling across a 90-min orbit with eclipse fraction.
• Thermal lag via first-order dynamics with distinct time constants.
• Task-driven CPU load spikes during payload ops and passes.
• Ground-pass windows with increased comm success rates.

• Battery undervoltage with negative slope and low SoC.
• Thermal overheating due to reduced radiative coupling or increased duty.
• Payload power surge causing 5 V rail droop.
• Communication blackout during scheduled pass.
• CPU hang and watchdog-triggered recovery.

• Eclipse drives energy budget; solar input zero in darkness.
• Thermal inertia causes lag between load changes and temperature.
• Pass geometry governs link availability and performance.
• Runaway processes and payload events mirror flight risks.

• Rule-based thresholds for immediate, explainable flags.
• Sliding-window trend analysis for slow drifts.
• Incremental statistics for mean, variance, and slope.

• Mean and variance via numerically stable updates.
• Linear slope from windowed sums for best-fit trend.
• Min/max and percentiles for bounds and spikes.

• Per-parameter scores normalized by thresholds.
• Subsystem health scores from weighted aggregation.
• Early warnings when sustained over limits.

• Track per-orbit V_batt minima and midday SoC maxima.
• Raise warning for >50 mV/orbit drop and >5% SoC decline.
• Escalate to critical for >100 mV/orbit or SoC <30% with nominal insolation.

• Power: undervoltage, low SoC, overcurrent.
• Thermal: overtemp, unsafe battery charge, rapid heating.
• Communication: blackout, persistent uplink failure.
• Software: CPU hang, memory leak, telemetry gap.
• Payload: overcurrent/overheat, non-responsive reset.

{
  "fault_type": "PowerFault",
  "severity": "Critical",
  "timestamp": "2026-01-05T12:34:56Z",
  "subsystem": "EPS"
}

• Nominal: full operations and standard thresholds.
• Limited: payload duty reduced, conservative limits, prioritized telemetry.
• Safe: payloads off, high-power rails off, beacon telemetry only.

• Power and thermal survival first.
• Communication continuity second.
• Payload operation last.

• Payload shedding and rail isolation.
• Subsystem restarts, escalating to OBC reboot.
• Safe-mode entry with beacon telemetry.

• Debounce entries, require persistence beyond short spikes.
• Clear exit with margins and minimum durations.
• Rate-limit restarts and escalation to avoid oscillation.

• Single fault: undervoltage during eclipse.
• Cascading faults: payload surge, bus droop, CPU spike, comm degradation.
• Long-term degradation: capacity loss across multiple orbits.
• False-positive avoidance: short comm fades and thermal spikes.

• Detection latency: Critical <10 s, Warning <60 s.
• Classification accuracy: Critical precision >95%, recall >85%.
• Survival time across eclipses under degraded EPS.
• Power saved via payload shedding and duty cycling.

• Deterministic seeds and scenario scripts at 1–2 Hz.
• Logs of telemetry, health scores, and fault events to ring buffers.
• Post-run metrics and parameter sweeps for robustness.

• Tight power and thermal budgets; intermittent communications.
• Delayed ground intervention requires on-board autonomy.

• Explainable, lightweight monitoring and recovery aligned to NanoSat constraints.
• Deterministic simulator for validation of nominal and faulted operations.

• Telemetry ingest, monitoring, classification, decision/recovery, logging/downlink.
• Simulator generates nominal and injected faults for repeatable tests.

• Incremental statistics for low CPU/memory.
• Hysteresis and rate limits to prevent oscillation.
• Safety hierarchy and mode transitions protecting spacecraft.

• Critical detection latency <10 s; Warning <60 s.
• Critical precision >95%; recall >85% across scenarios.
• Improved eclipse survival under degraded EPS.
• Quantified energy savings via payload shedding.

• Lightweight ML anomaly detectors trained offline, deployed on-board.
• Explainability preserved by combining ML scores with rule/trend evidence.

• Power and thermal survival.
• Communication continuity.
• Payload operation.

• Critical EPS/Thermal persists ≥10 s → Safe.
• Clear exit: V_batt ≥7.0 V and SoC ≥40% for ≥5 min, no active Critical faults.

• Process restart → Subsystem re-init → OBC reboot if persistent.
• Rate-limit: ≤1 restart per 5 min; 2 retries then escalate.
• Beacon telemetry when in Safe with summarized status.

• In-pass blackout: lower bitrate or increase FEC if available.
• Radio re-init and emergency telemetry prioritization.