Path Planning of Intelligent New Energy Vehicles Based on DQN

Path planning of intelligent new energy vehicles requires comprehensive consideration of multiple challenges such as dynamic traffic environment, energy efficiency and real-time decision-making. To solve the problem of insufficient adaptability of traditional optimization algorithms in dynamic scenarios, this paper proposes a path planning framework based on Deep Q Network (DQN). By fusing the perceptual power of deep learning with the decision-making mechanism of reinforcement learning, the model maps multidimensional state information into dynamic path decisions, and optimizes the training process by using experience playback, target network and ε-greedy policy. Simulation experiments show that DQN algorithm is significantly better than particle swarm optimization (PSO) and ant colony optimization (ACO) in path length (8.9% lower than traditional methods), running time (18.6%-31.1% higher) and optimal solution probability (98.1%), especially in obstacle dense scenes, showing stronger robustness and real-time performance, providing an efficient solution for autonomous decision-making of intelligent new energy vehicles.