Telemetry Logging

Telemetry logging is the practice of recording sensor data, state variables, and events during system operation for later analysis, debugging, and performance evaluation.

In AEGIS

AEGIS uses a dual-file logging strategy (ADR-013) to handle high-frequency data without blocking the 50Hz control loop:

CSV File (telemetry.csv)

• Format: Dense per-tick state snapshots
• Content: Position, velocity, attitude, engine throttles, FDI status, guidance commands
• Use: Post-flight analysis, trajectory replay, parameter tuning

JSONL File (events.jsonl)

• Format: Discrete event log (one JSON object per line)
• Content: State transitions, fault detections, dt spikes, aborts
• Use: Timeline of significant events, debugging

Implementation Details

• Buffered I/O: 1MB buffer to prevent blocking
• Symlink: logs/latest/ points to most recent run
• File rotation: Automatic based on timestamp

Post-Flight Analysis

# Trajectory analysis
python scripts/trajectory_analysis.py
 
# Attitude analysis
python scripts/attitude_analysis.py
 
# Dashboard
python scripts/trial_dashboard.py

Why Dual-Format?

Format	Pros	Cons
CSV	Easy to load, plot, analyze	Verbose for sparse events
JSONL	Flexible schema, event-based	Harder to query
Both	Best of both worlds	Slightly more code

Related Concepts

• Data Logging — General practice
• Telemetry — Remote data transmission
• Flight Data Recorder — Aviation equivalent (black box)
• Time Series Analysis — Analyzing logged data
• Digital Twin — Using logs to validate simulation models

Sources

• AEGIS Project (src/telemetry/, docs/data_flow_telemetry.md)