Deep Learning · Predictive Maintenance

UAV PredictiveMaintenance

An AI-driven Condition-Based Maintenance Plus (CBM+) digital twin. LSTM, Transformer, and 1D-CNN models ingest vibration, thermal, and power telemetry to forecast Remaining Useful Life across a tactical fleet — trained on the NASA C-MAPSS turbofan benchmark and a synthesized multirotor UAV fleet.

Launch the mission →See the models

01 / Results

What themodels achieved

Transformer14.12RMSE on held-out NASA C-MAPSS engines
UAV fleet97.4%Fault-classification accuracy
LSTM · Transformer · 1D-CNN3Architectures compared

02 / Explore

Inside thedigital twin

01 · Fleet Inventory · Asset Selection

The Armory

Interactive 3D carousel of all UAV and missile platforms. Inspect health, Remaining Useful Life, and select an asset to deploy.

Explore →
02 · Live Deployment · Real-Time Telemetry

The Mission

Tactical view with live radar sweep, streaming sensor telemetry, and dynamic Remaining Useful Life forecasting.

Explore →
03 · Post-Flight Diagnostics · XAI Overlay

The Digital Twin

Interactive 3D wireframe with anomaly highlighting, Transformer attention heatmap, and prescriptive maintenance recommendations.

Explore →
04 · LSTM × Transformer × 1D-CNN · C-MAPSS + UAV

The Model Lab

Architecture comparison with RMSE, PHM 2008 score, training curves, and per-asset prediction overlays.

Explore →