CompMobilSim:-- (Competitive Mobility Systems Simulator)

• 

  CompMobilSim Official Logo: Real-time Multi-Agent Simulation Engine

• 

  CompMobilSim Live Dashboard: Real-time Leaderboard, Track Minimap, and Telemetry Analytics

• 

  CompMobilSim 2D Track Visualization: Multi-Agent Movement in a Dynamic Urban Environment

• 

  CompMobilSim Telemetry Analytics: Agent Performance, Battery, and Collision Metrics

Inspiration

The rise of electric racing, drone competitions, and autonomous delivery systems inspired us to create CompMobilSim — a lightweight, real-time simulation tool that helps organizers, developers, and researchers analyze mobility events, optimize AI agents, and test strategies before real-world execution.

What it does

CompMobilSim is a modular, multi-agent simulation engine that:

Models cars, bikes, and drones with configurable physics and AI behavior

Simulates laps, checkpoints, pit stops, collisions, and environmental effects

Provides live leaderboards and telemetry analytics (speed, battery, position)

Allows users to define scenarios in JSON/YAML for reproducible simulations

Supports custom AI agent plugins via Python SDK It’s useful for racing strategy, drone competitions, logistics simulations, and urban mobility planning.

How we built it

Core Engine: Python (tick-based simulation, deterministic RNG)

Agent AI API: Python SDK for integrating custom behaviors

Frontend: React + Tailwind CSS for 2D interactive dashboard

Real-time Communication: FastAPI + WebSocket for live leaderboard updates

Storage & Analytics: SQLite and Pandas for telemetry logging

Scenario Composer: JSON/YAML configuration for track, agents, and events

Challenges we ran into

Ensuring deterministic, reproducible simulation runs with the same seed

Handling hundreds of agents in real-time efficiently

Designing a plugin API that is simple but flexible for AI agents

Integrating live leaderboard and telemetry visualization with minimal latency

Balancing simulation accuracy vs performance for scalable runs

Accomplishments that we're proud of

Built a fully modular, real-time multi-agent engine from scratch

Created a Python SDK enabling easy AI agent integration

Developed a live leaderboard and 2D visualization dashboard

Made simulations reproducible and configurable via JSON/YAML scenarios

Demonstrated applicability across racing, drones, and logistics planning

What we learned

Designing a deterministic simulation engine is critical for testing strategies

Modular APIs allow rapid integration of custom AI agents

Real-time visualization and telemetry require careful performance optimization

Multi-agent systems provide insights into strategy, behavior, and decision-making

Importance of balancing accuracy, scalability, and usability

What's next for CompMobilSim:-- (Competitive Mobility Systems Simulator)

Add 3D visualization for immersive analysis

Integrate reinforcement learning AI agents for autonomous racing

Support distributed simulations for large-scale logistics scenarios

Release as an open-source toolkit for researchers and mobility startups

Expand scenario library for different vehicle types, weather, and track conditions

Built With

• agent
• amazon-web-services
• d3.js
• docker
• ec2
• fastapi
• flask
• grafana
• json
• nginx
• numpy
• oauth2
• pandas
• plotly.js
• postgresql
• postman
• prometheus
• pytest
• python
• react
• redis
• socket.io
• sqlite
• timescaledb
• websockets
• yaml