摘要
针对智能车辆行驶环境和表达方式复杂等问题,提出了融合LSTM预测模型的智能车辆预测风险场模型。在传统势场模型基础上,考虑了动态目标行为预测信息,建立动态预测动能场,同时与道路环境中其他风险元素风险场相叠加,构建“预测风险场”统一模型。通过设计风险场代价函数完成规划轨迹簇的最小代价评估,获得最优路径规划轨迹。为验证该方法的有效性,进行了联合仿真和实车验证。结果表明,预测风险场模型有效表达了复杂行驶环境的交通态势,选取的最优路径提升了其综合安全性。
In view of complex driving environments and expression methods of intelligent vehicles,a predictive risk field model for intelligent vehicles incorporating long short term memory(LSTM)predic-tion model was proposed.Based on the traditional potential field model,the predictive information about dynamic target behaviors was considered,and the kinetic energy field of dynamic prediction was established.In addition,the kinetic energy field was superimposed with the risk field of other risk ele-ments in the road environment to construct a unified model,namely the predictive risk field.By design-ing the cost function of the risk field,the minimum cost evaluation of the planning trajectory cluster was completed,and the optimal path planning trajectory was obtained.In order to verify the effectiveness of the method,joint simulations and real-vehicle verification were conducted.The experimental results show that the predictive risk field model effectively expresses the traffic situation in the complex driving environment,and the selected optimal paths improve comprehensive safety.
作者
杨正才
谷师锐
吴浩然
孙文
Yang Zhengcai;Gu Shirui;Wu Haoran;Sun Wen(School of Automotive Engineering,Hubei University of Automotive Technology,Shiyan 442002,China;Hubei Key Laboratory of Automotive Power Train and Electronic Control,Shiyan 442002,China;School of Automotive Engineering,Changzhou Institute of Technology,Changzhou 213032,China)
出处
《湖北汽车工业学院学报》
2024年第1期7-12,17,共7页
Journal of Hubei University Of Automotive Technology
基金
中央引导地方科技发展专项(2022BGE248)
湖北汽车工业学院博士科研启动基金(BK202215)。
关键词
路径规划
轨迹预测
LSTM
预测风险场
path planning
trajectory prediction
LSTM
predicted risk field
