Skip to content

Train cars to drive using neural networks and evolve their performance through generations of evolution.

Notifications You must be signed in to change notification settings

mahdi-eth/Self-Driving-Car-Simulation

Repository files navigation

Self-Driving Car Simulation

A browser-based simulation training autonomous cars using a genetic algorithm. Cars equipped with neural networks learn to drive through generations of evolution. Sensors provide environmental data for decision-making. Train cars to navigate and complete driving tasks. Reinforcement learning and evolutionary computation combined in an interactive simulation.

Features

  • Genetic algorithm for evolving car driving behavior
  • Neural networks as car "brains" to control driving
  • Simulation of car-environment interaction
  • Evaluation metrics for measuring driving performance
  • Visualization of the neural network
  • Manual saving and discarding of the best performing car's brain