Model Engine Configuration
Physics-Guided ResNet Classifier (Built-in)
Drag and drop Pair 1's exported `.onnx` model file to evaluate real-time fleet inference, or use the pre-trained physics classifier.
3D Fleet Airspace Radar Dome
Multi-Class Aerial Swarm Tracking (Drones, Aircraft, Birds, Anomalies)
Filter Class:
Scale:
Drone Aircraft Bird Anomaly
Click any target on the 3D Radar to inspect real-time features, SHAP values, and threat assessment.
Active Fleet Target List
Confirmed Tracks
—
Tentative Tracks
—
Swarm Coherence
—
Swarm Threat
—
Model Performance Dashboard & Tactical Confusion Matrix
Total Swarm Sweeps0
Correct Classifications0
Overall Swarm Accuracy0.0%
Avg Inference Latency0.04 ms
Per-Class Validation Metrics
| Class | Swept | Correct | Recall % | Precision % |
|---|---|---|---|---|
| bird | 0 | 0 | 0.0% | 0.0% |
| drone | 0 | 0 | 0.0% | 0.0% |
| aircraft | 0 | 0 | 0.0% | 0.0% |
| anomaly | 0 | 0 | 0.0% | 0.0% |
Tactical Confusion Matrix (Actual vs Pred)
PREDICTED CLASS →
| ACTUAL ↓ | bir | dro | air | ano |
|---|---|---|---|---|
| bird | 0 | 0 | 0 | 0 |
| drone | 0 | 0 | 0 | 0 |
| aircraft | 0 | 0 | 0 | 0 |
| anomaly | 0 | 0 | 0 | 0 |
Fleet Stress Test Benchmark (Validated Backend Output)
Loading benchmark report…