Implementasi Deep Q-Learning dalam Pembuatan Model Self-Driving Car

Abstract

Abstract.

The development of self-driving cars has become a primary focus in artificial intelligence research. In this article, I introduce an approach using Deep Q-Learning and Artificial Neural Networks to develop a self-driving car model that can learn autonomously without predefined rules. The implementation of this approach utilizes Python version 3.6 programming language, the PyTorch library, and the Kivy framework to create a simulation environment for the self-driving car model. The self-driving car model is trained in a simulation that reflects real-world situations on the road. The evaluation results demonstrate the model's ability to recognize roads, avoid obstacles, and make intelligent decisions based on its surrounding environment. The implications of this research are the potential to develop smarter, adaptive, and safer self-driving cars in the future. The Deep Q-Learning method provides a solid foundation for self-driving cars to learn autonomously through experiences in various environmental situations. The research findings indicate that the implementation of Deep Q-Learning opens significant opportunities to create more advanced, intelligent, and autonomous self-driving cars in the future.
Keywords: Self-driving car, Deep Q-Learning, Neural Network, simulation, Python 3.6, PyTorch, Kivy.

Abstrak

Pengembangan self-driving car telah menjadi fokus utama dalam penelitian kecerdasan buatan. Dalam artikel ini, saya memperkenalkan pendekatan menggunakan teknik Deep Q-Learning dan Jaringan Saraf Tiruan untuk mengembangkan model self-driving car yang mampu belajar sendiri tanpa perlu aturan sebelumnya. Implementasi pendekatan ini menggunakan bahasa pemrograman Python versi 3.6, library PyTorch serta framework Kivy untuk membuat lingkungan simulasi model self-driving car. Model self-driving car ini dilatih dalam simulasi yang mencerminkan situasi nyata di jalan raya. Hasil evaluasi menunjukkan kemampuan model dalam mengenali jalan, menghindari rintangan, dan membuat keputusan cerdas berdasarkan lingkungan sekitarnya. Implikasi dari penelitian ini adalah potensi untuk mengembangkan self-driving car yang lebih pintar, adaptif, dan aman di masa depan. Metode Deep Q-Learning memberikan fondasi yang solid bagi self-driving car yang dapat belajar secara mandiri melalui pengalaman di berbagai situasi lingkungan. Hasil penelitian menunjukkan bahwa implementasi Deep Q-Learning membuka peluang besar dalam menciptakan self-driving car yang lebih maju, cerdas, dan mandiri di masa depan.

Kata Kunci: Self-driving car, Deep Q-Learning, Jaringan Saraf Tiruan, simulasi, Python 3.6, PyTorch, Kivy.