User Guide Overview

Welcome to the ROBIN user guide.

What ROBIN Provides

ROBIN ships two reusable modules and a documented integration pattern:

  • Module 1 - Process Intelligence API: process lifecycle, deviation detection, AI recommendations

  • Module 2 - Monitoring Dashboard: configurable operator UI with live KPIs and alerts

  • Integration pattern - ROS 2 to FIWARE: telemetry aggregation and DDS bridge config for connecting any ROS 2 robot to the FIWARE data layer

The components are domain-agnostic and demonstrated across welding and spray coating profiles.

Core Concepts

Entities in Orion-LD:

  • Process: operation mode, parameters, tolerance, state

  • Measurement: time-stamped measured values

  • GeometryTarget: target dimensions for geometry-driven operation

  • AIRecommendation: AI model output and guidance

Operation Modes

Parameter-driven

  • operator sets process parameters

  • AI predicts expected geometry from those parameters

  • telemetry is streamed

  • alert engine compares measured geometry against AI prediction

  • warnings/alerts are generated on true deviation beyond tolerance

Geometry-driven

  • operator sets target geometry

  • AI suggests process parameters for that target

  • telemetry is streamed

  • alert engine checks measured geometry against expected geometry from AI-guided settings

  • deviation checks in this mode use the fixed AI-guided/setpoint parameters rather than per-sample live measured parameters

  • reason: geometry deviation should be evaluated against the chosen plan (target + AI recommendation), without coupling the metric to control-loop jitter or parameter noise

  • warnings/alerts are generated on true deviation beyond tolerance

Daily Workflow

        graph LR
    START["docker compose up -d"] --> CREATE["Create / select\na process"]
    CREATE --> STREAM["Stream measurements\n(CLI or ROS 2 DDS)"]
    STREAM --> MONITOR["Monitor in\nROBIN Dashboard"]
    MONITOR --> DECIDE{"Deviation\ndetected?"}
    DECIDE -->|yes| REC["Review AI\nrecommendation"]
    REC --> APPLY["Apply / reject\nparameter change"]
    APPLY --> STREAM
    DECIDE -->|no| STREAM

  1. Start stack (docker compose up -d)

  2. Create/select a process

  3. Stream measurements (CLI or ROS 2 DDS path)

  4. Monitor deviations and alerts in the ROBIN Dashboard

  5. Apply recommendations and iterate

Interfaces

  • ROBIN Dashboard (http://localhost:5174): operator cockpit

  • Alert Engine API (http://localhost:8001/docs): integration and automation

  • Orion-LD API (http://localhost:1026/ngsi-ld/v1): direct NGSI-LD access

  • Mintaka API (http://localhost:9090/temporal): temporal data queries

ROS 2 / DDS Path

The current ingestion baseline is DDS-first:

  • ROS topics are aggregated into /robin/telemetry

  • Orion DDS mapping in config-dds.json writes into urn:robin:processTelemetry

Legacy HTTP NGSI bridge is intentionally removed from the recommended workflow.

Profiles

The core is domain-agnostic. Domain-specific demos are provided as profiles:

  • Welding (reference): python demo/profiles/welding_profile.py --mode both

  • Spray Coating: ROBIN_PROFILE=spray_coating docker compose up -d then python demo/profiles/spray_coating_profile.py --mode both

Each demo waits for you to press Start from the dashboard before streaming.

Each profile maps domain terms onto the same five core measurement fields (measuredHeight, measuredWidth, measuredSpeed, measuredCurrent, measuredVoltage). See demo/profiles/README.md for the comparison table and instructions for creating new profiles.

For publication details, see:

  • ARISE_PUBLICATION_ROADMAP.md

  • arise/catalog-metadata.yaml