Self-parking car using genetic algorithm (2021)

A self-parking car was developed using a genetic algorithm. The car is a toy robot car with a Raspberry Pi and a camera. The algorithm was trained on 1000 simulations with 1000 generations. The car was able to park in 90% of the simulations. The algorithm used a population size of 100 and a mutation rate of 0.1. The car's movement was simulated using the OpenCV library and the car's position was represented as a 2D point. The algorithm used a fitness function that penalized the car for colliding with the wall or other obstacles, and rewarded the car for parking within a certain distance from the target parking spot.