智能网联汽车多目标预测优化换道决策方法OA北大核心CSTPCD
Multi-Objective Predictive Optimization Based Lane Change Decision Making Method for Automated and Connected Vehicles
换道决策是智能网联汽车的核心难题之一,其面临着高动态、复杂交通场景下需要综合考虑行驶安全及效率等目标的巨大挑战.提出一种多目标预测优化的换道决策方法,主要包括动力学矩阵建模及多目标预测优化问题解算.基于智能网联汽车的通讯大数据信息构建交通流矩阵模型,然后分别设计表征车辆换道安全、行驶效率的动力学模型,通过多目标综合预测优化方法,求解条件约束下预测优化问题从而优化出最优换道决策指令.结果表明,所提出的预测优化换道方法较其他方法提高了智能汽车的行驶安全性和效率.
Lane change decision-making is one of the current opening challenges of automated and connected vehicles.Due to highly dynamic and complex traffic situations,multi-objective decision-making considering vehicle safety and riding efficiency is much more challenging.Therefore,this paper proposes a novel multi-objective predictive optimization-based lane change decision making method,which consists of dynamic matrix modeling and resolving of multi-objective predictive optimization problem.First,the matrix model of traffic flow is established based on the big data information from connected vehicles.Then,dynamic models representing lane change safety and riding efficiency are designed.The predictive optimization problem with constraints can be solved so that the optimal lane change decision is provided.Experimental results illustrate that the proposed method performs better and can improve vehicle safety and riding efficiency of automated and connected vehicles.
程硕;夏新;NAKANO Kimihiko
东京大学 生产技术研究所,东京 153-0041,日本加州大学洛杉矶分校 土木与环境工程系,洛杉矶 90095,美国
交通运输
智能网联汽车换道决策多目标优化预测优化
automated and connected vehicleslane change decision makingmulti-objective optimizationpredictive optimization
《同济大学学报（自然科学版）》 2024 (007)
1109-1117 / 9
日本学术振兴会外籍特别研究员资助项目(P21362)
评论