| 唐浩东,潘树国,陶贤露,高旺,张驰.基于CPO-LSTM神经网络的车辆运动状态分层式估计方法[J].电子测量与仪器学报,2026,40(6):150-158 |
| 基于CPO-LSTM神经网络的车辆运动状态分层式估计方法 |
| Vehicle motion state stratified estimation based on CPO-LSTM |
| |
| DOI： |
| 中文关键词: 车辆状态估计 动力学模型 机器学习 长短时记忆网络 冠豪猪优化算法 |
| 英文关键词:vehicle state estimation dynamic model machine learning long short-term memory (LSTM) crested porcupine optimizer |
| 基金项目:国家自然科学基金（42404023）项目资助 |
|
|
| Author | Institution |
| Tang Haodong | State Key Laboratory of comprehensive PNT Network and Equipment Technology, School of Instrument Science and
Engineering, Southeast University, Nanjing 210046, China |
| Pan Shuguo | State Key Laboratory of comprehensive PNT Network and Equipment Technology, School of Instrument Science and
Engineering, Southeast University, Nanjing 210046, China |
| Tao Xianlu | State Key Laboratory of comprehensive PNT Network and Equipment Technology, School of Instrument Science and
Engineering, Southeast University, Nanjing 210046, China |
| Gao Wang | State Key Laboratory of comprehensive PNT Network and Equipment Technology, School of Instrument Science and
Engineering, Southeast University, Nanjing 210046, China |
| Zhang Chi | State Key Laboratory of comprehensive PNT Network and Equipment Technology, School of Instrument Science and
Engineering, Southeast University, Nanjing 210046, China |
|
| 摘要点击次数: 64 |
| 全文下载次数: 10 |
| 中文摘要: |
| 准确获取车辆的运动状态对于车辆轨迹位置的推算具有重要意义,尤其是在卫星信号遮蔽严重的环境中(如城市峡谷、树林、隧道等场景),而现有的基于动力学模型驱动的车辆运动估计方法过于依赖建模精度,且在一些极限工况下难以进行准确的估计。故为了更加简洁准确地获取车辆运动状态信息,提出了一种基于冠豪猪算法优化的长短时记忆网络（CPO-LSTM）的分层式车辆状态估计方法。首先,通过分析车辆动力学基本物理特性,设计了纵向速度与侧向速度的分层式估计策略；然后，构建多种行驶工况组成的数据集进行训练,在训练过程中利用冠豪猪优化算法（CPO）对学习率与迭代次数等参数进行寻优;最后,对估计方法进行了多工况及多路面条件下的虚拟仿真试验验证。结果表明,该方法有效地提高了车辆运动状态估计精度,横向速度、纵向速度和横摆角速度的估计精度相比于基于模型驱动的方法精度分别提升92%,73%和52%,整体表现优于传统的基于动力学模型的卡尔曼滤波算法。 |
| 英文摘要: |
| Accurately obtaining the motion state of a vehicle is of great significance for the estimation of vehicle trajectory and position, especially in environments with severe satellite signal occlusion (such as urban canyons, wooded areas, tunnels, etc.). However, existing vehicle motion estimation methods driven by dynamic models rely too much on modeling accuracy and are difficult to achieve accurate estimation under some extreme working conditions. Therefore, in order to accurately obtain vehicle motion state information without relying on the accuracy of dynamic models, this paper proposes a hierarchical vehicle state estimation method based on the crested porcupine optimization long short-term memory network. Firstly, by analyzing the basic physical characteristics of vehicle dynamics, a hierarchical estimation strategy for longitudinal velocity and lateral velocity is designed. Then, a dataset composed of various driving conditions is constructed for training, and during the training process, the crested porcupine optimization algorithm is used to optimize parameters such as learning rate and number of iterations. Finally, virtual simulation tests are conducted to verify the estimation method under multiple working conditions and multiple road surface conditions. The results show that this method effectively improves the estimation accuracy. Compared with model-driven methods, the estimation accuracy of lateral velocity, longitudinal velocity, and yaw rate is improved by 92%, 73%, and 52% respectively, and the overall performance is better than the traditional Kalman filtering algorithm based on dynamic models. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
