PhyTrace¶

PhyTrace is the foundational data collection layer for autonomous systems. It captures comprehensive telemetry and trace data from robots and streams it to PhyCloud for storage and analysis.

Overview¶

graph LR
    A[Robot] --> B[PhyTrace Agent]
    B --> C[UDM Normalization]
    C --> D[Transport Layer]
    D --> E[PhyCloud]
    E --> F[PhySafe]
    E --> G[PhyComp]
    E --> H[Phylyze]

SDKs¶

PhyTrace provides SDK libraries in multiple languages:

Key Concepts¶

Unified Data Model (UDM)¶

All telemetry is normalized to the PhyTrace UDM — a standardized schema covering 23 domains (location, motion, power, safety, AI reasoning, etc.). This ensures platform-independent analytics regardless of robot type.

Platform Adapters¶

Adapters bridge platform-specific APIs (ROS2, Tesla Fleet API, custom REST, etc.) to the UDM. The SDK handles the translation automatically.

Provenance¶

Every event is tagged with cryptographic provenance metadata — source, timestamp, chain of custody — enabling trust verification from capture to cloud.

Architecture¶

┌────────────────────────────────────────────────────┐
│                  PhyTrace Agent                    │
│                                                    │
│  Platform Adapter → UDM Builder → Validation       │
│       → Provenance Tagger → Buffer/Batch           │
│       → Transport (HTTP/gRPC/MQTT/WS)              │
└────────────────────────────────────────────────────┘