Self-parking car using genetic algorithm (2021)

Researchers developed a self-parking car simulation using a genetic algorithm, where virtual cars with random genomes gradually learn to navigate into parking spots over successive generations. The system equips simulated cars with "muscles" (engine and steering controls), "eyes" (sensors to detect obstacles), and a "brain" (a function that translates sensor input into movements), which evolves through natural selection principles. By the 40th generation, the cars demonstrate significant improvement in parking behavior, and users can interact with a browser-based simulator to train cars from scratch or observe the evolved self-parking in action.