Fast and curious

Race. Learn. Adapt. Repeat.

Comment

ppt link : https://shorturl.at/uer9v

What It Does

The Competitive Mobility Systems Simulator is an agent-based, real-time simulation platform that models multiple mobile entities — cars, drones, and human agents — competing across dynamic environments.
Each agent operates with unique decision logic, responding to traffic, obstacles, and environmental events (like weather or accidents).

The system supports:

  • Real-time visualization on a live interactive map.
  • Leaderboard and scoring system based on performance metrics like speed, efficiency, and route optimization.
  • Configurable simulation scenarios using map data from OpenStreetMap.
  • IoT integration through MQTT for real-time telemetry from physical devices (digital twins).

How We Built It

Component Description Technologies
Simulation Engine Agent-based modeling and event-driven simulation Python • Mesa • FastAPI
Backend Server Real-time WebSocket and API handling FastAPI • Redis (Pub/Sub)
Frontend Visualization Live map rendering and leaderboard React • Mapbox GL JS
IoT Gateway Telemetry ingestion from sensors and devices MQTT • Eclipse Mosquitto • ESP32 / Arduino
Database Layer Persistent storage for scenarios, metrics, and events PostgreSQL • TimescaleDB
Containerization Multi-service orchestration for reproducible setup Docker • Docker Compose

Challenges We Ran Into

  • Balancing real-time performance with simulation complexity for multiple concurrent agents.
  • Creating efficient state synchronization between backend simulation and frontend visualization.
  • Integrating live IoT telemetry with simulated digital twins in real time.
  • Managing scalability for hundreds of agents without latency or performance drops.
  • Designing a clean and responsive UX for live visualization and competition tracking.

Accomplishments That We’re Proud Of

  • Achieved real-time simulation visualization with dynamic leaderboard updates.
  • Successfully integrated MQTT-based IoT devices streaming telemetry into the simulation.
  • Built a configurable simulation engine allowing user-defined scenarios, agents, and competition rules.
  • Designed a modular architecture supporting future extensions (e.g., drones, delivery bots).

What We Learned

  • Importance of modular and asynchronous architectures for real-time simulation systems.
  • How to use digital twin principles to link physical IoT data and virtual agents effectively.
  • Techniques for optimizing Python event loops and WebSocket throttling for smoother updates.
  • Gained insights into mobility system dynamics, traffic behavior, and adaptive routing logic.

What’s Next for the Project

  • Implement reinforcement learning-based agents for adaptive decision-making.
  • Expand IoT integration to include real drones and autonomous vehicles.
  • Add 3D visualization using Three.js or Unreal Engine for immersive simulation.
  • Develop distributed simulation clusters to scale thousands of concurrent agents.
  • Build a scenario marketplace where users can upload and share competitive mobility challenges.

Built With

Share this project:

Updates