Adaptive Cruise Control Algorithm for Intelligent Vehicles in Highway Environments Based on Deep Q-Network

This paper explores the application of the Deep Q-Network (DQN) algorithm in addressing the challenges associated with adaptive cruise control systems for modern automobiles. To achieve this, a reinforcement learning model has been developed to simulate vehicle dynamics, wherein the algorithm processes the vehicle’s current state to predict and determine the most suitable control actions. The design of the reward function incorporates multiple constraints, including real-time velocity and both upper and lower speed thresholds, to ensure the agent makes sound decisions during operation. To assess the algorithm’s performance in a realistic context, a simulated highway environment was created, and extensive experiments were conducted. The results from these simulations indicate that utilizing the DQN algorithm for automotive cruise control not only decreases computational demands but also enhances the safety and stability of the vehicle’s movement on the road.