Independent evaluation system for UAV BVLOS flight performance — aligned with the EU-funded EARHEART project on cloud-based autonomous drone operations.
This tool provides structured, evidence-based evaluation of BVLOS drone flight telemetry data against EASA SORA operational KPIs. It mirrors the independent evaluation responsibilities of BCU on the EARHEART project — assessing AI-enabled drone platforms, evaluating sensor performance against defined trial KPIs, and producing structured outputs for regulatory and consortium reporting.
Core capabilities:
- Real-time KPI evaluation against 7 EASA SORA operational standards
- Automated anomaly detection and event logging
- Sensor performance visualisation (GPS, C2 link, battery, altitude, speed, wind)
- Interactive dark-theme mission control dashboard
- Structured JSON report export for consortium and regulatory submission
- Live flight path mapping with altitude colour coding
| KPI | Threshold | Regulatory Basis |
|---|---|---|
| GPS Positioning Accuracy | HDOP < 2.0 | EASA AMC RPAS.1309 |
| GPS Satellite Coverage | Min 6 satellites | UK CAA CAP 722 |
| C2 Link Signal Strength | > -75 dBm | EASA SORA OSO #10 |
| Battery Reserve Margin | Never below 20% | EASA UAS.SPEC.050 |
| Altitude Holding Accuracy | Deviation < 5m | EASA SORA OSO #06 |
| Speed Limit Compliance | Max 25 m/s | EASA ConOps Envelope |
| Wind Envelope Compliance | Max 10 m/s | Manufacturer limits |
- App: Streamlit (dark mission control theme)
- Visualisation: Plotly (interactive charts + radar)
- Data: Pandas, NumPy
- Risk Engine: Custom SORA-aligned KPI scoring
- Language: Python 3.10+
git clone https://github.com/Lakshan-D/drone-telemetry-analyser.git
cd drone-telemetry-analyserpython -m venv venv
venv\Scripts\activate # Windows
source venv/bin/activate # Mac/Linuxpip install -r requirements.txtpython generate_sample_data.pyCreates three sample flight logs in data/sample_logs/ — one nominal flight and one with injected anomalies for demonstration.
streamlit run app.pyOpens at http://localhost:8501
drone-telemetry-analyser/
├── app.py # Main Streamlit application
├── generate_sample_data.py # Sample telemetry generator
├── requirements.txt
├── utils/
│ └── kpi_engine.py # SORA-aligned KPI evaluation engine
└── data/
└── sample_logs/ # Generated flight telemetry CSVs
timestamp, latitude, longitude, altitude_m, speed_ms, battery_pct,
gps_satellites, hdop, signal_dbm, pitch_deg, roll_deg, yaw_deg,
wind_speed_ms, temperature_c
Compatible with DJI, ArduPilot, and PX4 log exports (with column renaming).
The EARHEART EU project trials cloud-based BVLOS drone operations and produces regulatory outputs including a BVLOS White Paper. BCU's role is independent evaluation of trial platforms. This tool demonstrates the core evaluation methodology:
Flight telemetry → KPI assessment → Anomaly detection → Regulatory report → Policy recommendation
Lakshan Divakar MSc Electronics & Electrical Engineering, Brunel University London Research Lab Assistant — LiDAR, Drones, AGV, Sensor Fusion
MIT License — free to use, modify, and share with attribution.