Welcome to the Automatic Parking System project! This repository showcases an autonomous vehicle parking solution, utilizing Reinforcement Learning (RL) and Genetic Algorithms (GA), and validated through simulation in Unity.
This project aims to develop an autonomous parking system capable of performing various parking maneuvers like parallel, perpendicular, and angular parking. The system uses Reinforcement Learning to teach the agent optimal parking strategies in diverse urban scenarios.
Project Objectives:
- Develop a self-learning model for automated parking using RL.
- Test and validate the model in Unity for real-world applicability.
- Parking Strategy Optimization: The RL model learns efficient paths for safe parking maneuvers.
- Obstacle Avoidance: Ensures the vehicle avoids nearby obstacles during parking.
- Urban Environment Testing: Validates the model’s effectiveness across various urban parking scenarios.
- Unity: For creating and simulating complex parking environments.
- Reinforcement Learning: Trains the parking model to make efficient decisions.
- Genetic Algorithms: Optimizes model parameters for enhanced performance in challenging urban settings.
- Python: Used for the RL model and backend processing.
Check out a video demo of the model in action below:
video1276435030.mp4
After training, the model exhibits improved parking performance, achieving high success rates with reduced collision occurrences.
- Success Rate: 66%
- Average Time per Parking: 13 seconds
- Collision Rate: 34%
- Adapting to varied parking scenarios with limited sensor data.
- Minimizing training time while achieving optimal performance.
This project is licensed under the MIT License. See the LICENSE file for more details.
Let us know if you have any questions or ideas for improvements! 😊