AI Racing Simulator

Track Selection

â–ŧ

Stats

â–ŧ

Mode: Manual

Speed: 0 km/h

Lap Time: 0.00s

Record: --

Generation: --

Alive: --/--

Best Fitness: --

🧠 Neural Network

â–ŧ

Click đŸ§ đŸ‘ī¸ button in header, then:

â€ĸ Start AI Training (🧠) to see the leader's brain

â€ĸ Run Best Model (🏆) to see autopilot brain

🎉

Track Mastered!

Generation --
Laps Completed --
Best Fitness --

Your AI has successfully mastered this track! The best model has been saved automatically.

âš™ī¸

Training Settings

Training Parameters

More cars = better learning but slower (5-200)
How much genes change (0.05-0.20 recommended)
Max frames per generation (300-5000)

Visualization

Render every Nth generation (1-100, higher = faster)
Steps per update: 10 (1-20)

Auto-Stop

Stop training when track is mastered
Laps needed to consider track mastered (1-20)
Cars that must complete laps (1-50)
â„šī¸

AI Racing Simulator

🎮 Manual Controls

Arrow Keys: Drive the car manually

↑: Accelerate

↓: Brake/Reverse

←/→: Steer left/right

🧠 AI Training

Start Training: Click the brain icon (🧠) to let AI cars learn

How it Works: Uses genetic algorithm with neural networks. Cars evolve over generations to find the optimal racing strategy.

Auto-Stop: Training automatically stops when multiple cars master the track (configurable in settings)

🏆 Running Models

Test Best Model: Click trophy icon (🏆) to watch the best AI drive

Save Model: Download trained models with the save icon (💾)

Track Selection: Train different models on 5 difficulty levels

âš™ī¸ Settings

Population Size: More cars = better learning but slower

Mutation Rate: Controls genetic variation (0.05-0.20 recommended)

Simulation Speed: Adjust visual pacing during training

Render Interval: Skip generations for faster training

đŸ‘ī¸ View Toggle

Toggle View: Show/hide sensor rays and slow down training for detailed observation

🧠 Neural Network Visualization

View the AI Brain: Click the đŸ§ đŸ‘ī¸ button in the header to see how the AI thinks in real-time

What You'll See:

  • Input neurons showing sensor readings and speed
  • Hidden layers processing the information
  • Output neurons controlling steering and throttle
  • Connection weights (green = positive, red = negative)
  • Live activation values as the AI makes decisions

During Training: Shows the leader car's brain

During Autopilot: Shows the running model's brain

🏆

Select Model to Run

🧠

Current Best Model

Run the currently trained model

No model trained yet

OR
📁

Load Saved Model

Upload a previously saved model file