12 / 12 · Database systems
CubeSat Telemetry
Exploring telemetry data management for a self-healing satellite database through a synthetic-data case study.
Downlink · 6 channels
Break it. Watch it heal.
- Records
- 0
- Faults
- 0
- Repaired
- 0
- Quarantined
- 0
- Integrity
- 100%
| Seq | UTC | Bat V | Temp °C | Solar mA | Gyro °/s | Mag µT | RSSI | CRC32 | Quality |
|---|
- Category
- Database systems · Engineering
- Team
- Four members
- Engine
- SQLite · Python
- Data
- NASA ELFIN-B telemetry · synthetic faults
What it does
- A working prototype tested on real NASA ELFIN-B spacecraft telemetry with artificially injected corruption.
- SQLite with write-ahead logging, SHA-256 integrity checks, Isolation Forest anomaly detection and temporal-interpolation recovery — each repair is verified, versioned and logged with where it came from.
- On-board database design on SQLite: a 12-table schema in Third Normal Form.
- Per-record CRC32 checksums, write-ahead logging and trigger-based version history.
- Layered fault detection and interpolation-based recovery with quality flags.
- Telemetry processing evaluated on simulated six-channel data with injected faults.
- Limitations documented; replay on real flight data left as future work.
Radiation bit flips, sensor glitches and replayed frames quietly corrupt satellite data. The project asks whether standard database techniques, applied record by record, can make that data repair itself. The course case study was evaluated on synthetic data; the prototype runs on real ELFIN-B telemetry with injected faults.
Case study, no public link.