Self-Parking Car Evolution

Genetic algorithm demo that evolves virtual self-parking cars in the browser.

5.0 (4)

Geprüft von Daniel Nikulshyn·Aktualisiert Mai 2026

Neural Networks Research Open Source Educational Browser-Based Genetic Algorithm Simulation

Übersicht

Self-Parking Car Evolution is an open educational project that uses a genetic algorithm to teach simulated cars how to park themselves in a 2D virtual environment. Each car is controlled by a small neural network whose weights are encoded as a genome, and successive generations are bred, mutated, and selected based on how close they get to the target parking spot. The simulation runs entirely in the browser, letting users watch the population improve over time as poorly performing cars are filtered out and stronger drivers pass on their parameters. It serves as a hands-on illustration of evolutionary computation, fitness functions, and emergent behavior rather than a production-ready autonomous driving system. Developers, students, and AI enthusiasts can explore the source code to learn how genetic algorithms work in practice, tweak parameters, or adapt the approach to other control problems.

Hauptfunktionen

Genetic algorithm-based training loop
Neural network car controllers
2D parking simulation environment
Configurable population and mutation parameters
Live visualization of evolving generations
Open-source codebase for experimentation

Anwendungsfälle

Learn Genetic Algorithms Visually

Students and self-learners can watch populations of cars evolve in real time to build intuition about selection, mutation, and fitness functions.

Classroom Demo for Evolutionary AI

Instructors can use the in-browser simulation as a live teaching aid when introducing neuroevolution, emergent behavior, or reinforcement-style learning concepts.

Experiment with Hyperparameters

Developers can tweak population size, mutation rates, and network weights to study how these parameters affect convergence speed and parking success.

Starter Project for Neuroevolution

Hobbyists and researchers can fork the open-source codebase as a foundation for building their own genetic algorithm experiments or simulation environments.

Pro & Contra

Pro

Clear, visual demonstration of genetic algorithms
Runs in the browser with no setup
Open source and educational
Good entry point for evolutionary AI concepts

Contra

Limited to a toy parking scenario
Not suitable for real-world autonomous driving
Training can be slow to converge
Requires coding knowledge to extend

Bewertungen

5.0

Durchschnitt aus 4 Bewertungen.

Melde dich an, um eine Bewertung abzugeben.

Sanjay Gupta

Compared a few options

Evaluated this against two competitors. Where it wins: neural network car controllers and clear, visual demonstration of genetic algorithms. Where it lags: training can be slow to converge. On balance the feature set — especially genetic algorithm-based training loop — justifies the 5 stars for our use case.

Wei Chen

Does the job

Pretty happy overall. Open-source codebase for experimentation just works and clear, visual demonstration of genetic algorithms. but no dealbreakers — I'd recommend it to a friend without hesitating.

Tomáš Novák

Skeptical, then convinced

I went in skeptical — most tools in this space overpromise. It actually delivers on 2D parking simulation environment, and clear, visual demonstration of genetic algorithms caught me off guard. still, I'd recommend giving it a real trial.

Omar Haddad

Solid for our team

We rolled this out across the team last quarter and clear, visual demonstration of genetic algorithms. Neural network car controllers fits neatly into how we already work, and 2D parking simulation environment removed a step we used to do by hand. but it has held up under daily use.