2. Scope

This RFC covers:

• The standard export structure for RS3-generated datasets
• Mapping between RS3 simulation variables and Telemachus field groups
• Quality control and noise injection strategies
• Validation workflow to produce certified Telemachus-compatible datasets

It does not cover RS3’s internal physics or simulation engine, only its output interface.

3. Relationship to Other RFCs

4. Pipeline Overview

The RS3 → Telemachus pipeline follows four sequential stages:

graph TD
  A["RS3 Simulation Engine"] --> B["RS3 Exporter"]
  B --> C["Telemachus Mapper"]
  C --> D["Validator (telemachus‑py)"]
  D --> E["Dataset Publication (Zenodo & GitHub)"]

Fallback (text-only):

RS3 Simulation Engine -> RS3 Exporter -> Telemachus Mapper -> Validator (telemachus-py) -> Dataset Publication (Zenodo/GitHub)

1. Simulation Engine — Generates raw signals (IMU, GNSS, vehicle dynamics).
2. Exporter — Converts simulation logs into structured CSV or Parquet files.
3. Mapper — Translates RS3 variable names to Telemachus schema fields.
4. Validator — Runs consistency checks and produces a dataset manifest.

5. File and Directory Structure

rs3_out/
 ├── raw/
 │    ├── imu.csv
 │    ├── gps.csv
 │    ├── vehicle.csv
 │    └── env.csv
 ├── mapped/
 │    ├── samples.csv
 │    └── dataset.json
 ├── logs/
 │    └── rs3_run.log
 └── out/
      └── 2025-10-13-v0.2/
           ├── samples.csv
           └── dataset.json

6. RS3 to Telemachus Field Mapping

Mapping files are stored under:

rs3/mappings/telemachus_v0.2.yaml

7. Dataset Generation Workflow

1. Run RS3 simulation with configuration file:

   python -m rs3 run --config configs/urban_day.yaml --out rs3_out/

2. Convert simulation output to Telemachus:

   python -m rs3.export.telemachus --input rs3_out/raw --output rs3_out/mapped

3. Validate generated dataset:

   telemachus validate --dataset rs3_out/mapped --level strict

4. Publish validated dataset:

   telemachus dataset publish --input rs3_out/mapped --to zenodo

8. Quality Control and Noise Injection

RS3 supports several noise models that improve realism:

All parameters are stored in the dataset manifest under:

"simulation": {
  "noise_model": "gaussian",
  "noise_parameters": {"sigma_accel": 0.02, "sigma_gyro": 0.01}
}

9. Validation Rules

Datasets exported from RS3 must pass:

• JSON schema validation (RFC-0003)
• Alignment tolerance tests (RFC-0007)
• Sampling rate consistency (10 Hz ± 0.1 Hz)
• Metadata completeness (version, source, description, license)

If any check fails, a warning or error is raised in the validation report.

10. Example Dataset Manifest

{
  "version": "0.2",
  "schema": "https://telemachus3.github.io/telemachus-spec/schema/dataset.schema.json",
  "created_at": "2025-10-13T09:30:00Z",
  "records": 86400,
  "sampling_rate_hz": 10.0,
  "sources": ["RS3"],
  "description": "Urban driving scenario with synthetic IMU and GNSS data",
  "license": "CC-BY-4.0",
  "simulation": {
    "engine": "RS3 core2",
    "noise_model": "gaussian",
    "noise_parameters": {"sigma_accel": 0.02, "sigma_gyro": 0.01}
  },
  "contact": "sebastien.edet@telemachus.org"
}

11. Integration with Validation Tools

The RS3 pipeline integrates directly with telemachus-py:

• Exposes CLI commands for dataset export and validation
• Provides metadata enrichment (simulation block)
• Supports auto-checks for monotonic timestamps and alignment

Each validated dataset includes a checksum.md5 and a VALIDATED: true tag.

12. Publication and Reproducibility

Certified datasets can be published under the Telemachus Datasets organization:

• GitHub (telemachus3/telemachus-datasets)
• Zenodo with DOI and version tag
• Metadata aligned with dataset.json

Each dataset must include:

• RS3 configuration file (configs/*.yaml)
• Validation report (out/validation.json)
• License and citation info (LICENSE, CITATION.cff)

13. Future Extensions

• Multi-vehicle and traffic scenario simulation
• Real-time RS3 → Telemachus streaming adapter
• Integration with OpenTelemetry metrics (RFC-0014)
• Support for RS3 plugin ecosystem (altitude, weather, fleet)
• Hybrid datasets combining simulated + real GNSS data

14. References

• RFC-0001 — Telemachus Core 0.2
• RFC-0003 — Dataset Specification 0.2
• RFC-0004 — Extended FieldGroups Schema
• RFC-0007 — Validation Framework & CLI Rules
• RFC-0010 — Synthetic Data Quality & Noise Models
• RS3 Project — https://github.com/SebE585/RoadSimulator3
• Telemachus Datasets — https://github.com/telemachus3/telemachus-datasets

15. Conclusion

This RFC defines a reproducible integration pipeline between RS3 and Telemachus, enabling the automatic generation of validated, standard-compliant datasets.
By bridging physical simulation and open data standards, it ensures that synthetic data can serve as a scientific and industrial reference for mobility and telematics research.