STRATOS 🏎️

Industrial-Grade Formula 1 Race Strategy Simulation Engine

STRATOS is a high-performance race strategy decision-support platform for Formula 1. It ingests live telemetry from the OpenF1 API, models tyre degradation and vehicle performance, and runs massive Monte Carlo race simulations to recommend optimal pit timing via an interactive HUD-style dashboard.

---

🤖 Multi-Agent System (MAS) Architecture

STRATOS is powered by an autonomous event-driven Multi-Agent architecture that decouples data ingestion from strategy computation:

graph TD
    API[OpenF1 API] --> TA[Telemetry Agent]
    TA --> SM[State Manager]
    SM --> DA[Dashboard App]
    DA --> SA[Strategy Agent]
    SA --> SIA[Simulation Agent]
    SIA --> MC[Monte Carlo Engine]
    MC --> SA
    SA --> |Strategy Insight| DA

Loading

• Telemetry Agent: Handles asynchronous ingestion, rate-limiting, and state synchronization.
• Simulation Agent: Autonomously manages the heavy-duty physics rollouts and scenario analysis.
• Strategy Agent: Prescribes optimal pit windows and traffic risk assessments based on agent-coordinated insights.

---

✨ Features

• Live Leaderboard: Real-time interval tracking, stint history, and tyre compound visualization.
• Strategy HUD: Dynamic "Action Badges" (PIT/STAY) with predictive position gain and undercut success probability.
• Telemetry Explorer: Deep-dive lap time analysis with dirty-air (traffic) impact indicators.
• Track Map: Real-time GPS-derived car positioning on a specialized vector track map.
• Race Replay: Complete session play/pause/scrub capability to analyze previous laps after the race.

---

Quick Start

1. Clone & install

cd STRATOS
python -m venv .venv && source .venv/bin/activate
pip install -e ".[dev]"

2. Configure

cp .env.example .env
# Edit .env – set OPENF1_BASE_URL, POLL_INTERVAL_SECONDS, SIM_COUNT, etc.

Viewing Other Races

By default, STRATOS uses the session_key=latest parameter from the OpenF1 API, which currently points to the 2026 Australian Grand Prix (Melbourne).

To view data from a different race, set the SESSION_KEY environment variable before starting the dashboard.

Example Sessions (2024 Races)

• Bahrain GP: SESSION_KEY=9472
• Saudi Arabian GP: SESSION_KEY=9480
• Australian GP: SESSION_KEY=9488
• Japanese GP: SESSION_KEY=9496
• Monaco GP: SESSION_KEY=9523

# Example: Start dashboard with Bahrain 2024 data
SESSION_KEY=9472 python -m stratos.dashboard.app

3. Run tests

pytest tests/ -v

4. Launch the dashboard

python -m stratos.dashboard.app
# Open http://127.0.0.1:8050

---

Module Overview

Module	Description
stratos/agents/	Autonomous Multi-Agent System governing data flow and decisions
stratos/ingestion/	OpenF1 REST client + Pydantic schemas
stratos/state/	Race state dataclass + thread-safe state manager
stratos/models/	Tyre degradation, vehicle performance, traffic, pit stop
stratos/simulation/	Monte Carlo race simulator + safety car model
stratos/strategy/	Strategy evaluator, optimizer, recommendation output
stratos/dashboard/	Plotly Dash app with live telemetry and strategy panels

---

Data Sources

All telemetry is sourced from the OpenF1 public API (no authentication required):
https://api.openf1.org/v1

Endpoints used: /laps, /position, /pit, /car_data, /weather, /track_status

---

Configuration

Variable	Default	Description
OPENF1_BASE_URL	https://api.openf1.org/v1	API root
POLL_INTERVAL_SECONDS	3	Telemetry polling cadence
SIM_COUNT	10000	Monte Carlo trials per strategy
DEFAULT_PIT_LOSS_SECONDS	23.0	Pit time loss fallback
DASH_HOST	127.0.0.1	Dashboard host
DASH_PORT	8050	Dashboard port

---

Tech Stack

• Python 3.11+ · NumPy · Pandas · SciPy
• Plotly Dash + Dash Bootstrap Components
• HTTPX (async HTTP) · Pydantic v2 (data validation)
• pytest · respx (mock HTTP) for testing