基于多目标优化的车辆编队仿真研究

Research on Vehicle Formation Simulation Based on Multi Objective Optimization

摘要: 随着车联网技术的发展，出现了车路协同、多车协同等多项汽车智能化技术。智能网联汽车编队在城市道路场景下有着良好的应用前景，在城市道路场景下，路况复杂，对车辆编队的灵活性提出了很高的要求。为了提高在城市街道场景下车辆编队的灵活性，提出了一种分散运动控制方法。该方法实现了编队当中每一个车辆的轨迹的优化和车间距的控制。首先，提出了一种车辆轨迹的生成方法，这种方法将车辆的横向轨迹和纵向轨迹解耦。同时，采用五次多项式生成轨迹族，并使用多目标优化选择最优轨迹。其次，针对车间距和车速的控制，设计了PD控制器，使得编队当中的车间距和车速能够达到一致。最后，通过Matlab仿真验证了分散式运动规划方法在不同场景下的优越性。结果表明，该方法提高了编队的灵活性和一致性。

Abstract: With the development of vehicle networking technology, multiple intelligent technologies for automobiles have emerged, such as vehicle road collaboration and multi vehicle collaboration. Intelligent connected vehicle formations have good application prospects in urban road scenarios, where complex road conditions require high flexibility for vehicle formations. In order to improve the flexibility of vehicle formation in urban street scenes, a decentralized motion control method is proposed. This method optimizes the trajectory of each vehicle in the formation and controls the distance between vehicles. Firstly, a method for generating vehicle trajectories is proposed, which decouples the lateral and longitudinal trajectories of the vehicle. Meanwhile, a fifth degree polynomial is used to generate trajectory families, and multi-objective optimization is employed to select the optimal trajectory. Secondly, a PD controller was designed to control the distance and speed between vehicles in the formation, ensuring consistency between the distance and speed. Finally, the superiority of the decentralized motion planning method in different scenarios was verified through Matlab simulation. The results indicate that this method improves the flexibility and efficiency of the formation.