摘要
我国下一代列车运行控制中,列车位置主要通过融合以卫星定位为核心的多源定位信息得到。然而,卫星量测易受到钟跳或环境多源噪声等不确定因素影响,产生阶跃故障或瞬时粗大偏差等现象,导致传统扩展卡尔曼滤波估计失准且稳定性下降。通过分析常规滤波方法在列车位置估计中存在的问题,采用信息理论学习思想提出一种基于最大相关熵准则的自适应扩展卡尔曼滤波定位方法(AMCEKF);利用最大相关熵准则替换传统滤波中的最小均方差准则,选择高斯核函数作为代价函数,重构量测噪声,避免了量测噪声的先验高斯假设;设计基于Pseudo-Huber的核宽度自适应更新策略,解决核宽度对估计性能的制约问题。采用拉萨—林芝铁路实测数据进行测试,测试结果表明:在瞬时故障和阶跃故障场景下,AMCEKF均能够有效抑制故障量测带来的定位性能退化,具有较高的鲁棒性和估计精度;相比基于扩展卡尔曼滤波的定位结果,水平位置精度分别增加了56.5%和61.2%。
Train position will be obtained by fusing multi-source positioning information with satellite navigation as the core in the next-generation train operation control system. However, during the train operation, the satellite measurement that likely generates step faults or instantaneous large deviations affected by uncertain factors, such as clock jumps or environmental multi-source noise, gives rise to the estimate inaccuracies and lower stability of traditional extended Kalman Filter. By analyzing the problems of traditional filtering methods in train position estimation, this paper presented an adaptive Extended Kalman Filter positioning method based on Maximum Correntropy Criterion(AMCEKF). The Maximum Correntropy Criterion was introduced to replace the minimum mean square error criterion in the traditional filtering. Reorganizing the measurement noise by gaussian kernel used for cost function could avoid priori gaussian hypothesis for measuring noise. In order to further solve the estimation performance constraint caused by the kernel width, Pseudo-Huber-based adaptive update strategy was designed for adjusting the kernel width. The field data of the Lhasa—Nyingchi railway were used to verify the proposed method. The results demonstrate that the AMCEFK can effectively suppress the degraded positioning performance caused by fault measurements under the scenarios of transient faults or step faults, with high robustness and estimation accuracy. Compared with the EKF-based positioning results, the horizontal position accuracy is increased by 56.5% and 61.2% respectively.
作者
王韦舒
上官伟
刘江
姜维
WANG Weishu;SHANGGUAN Wei;LIU Jiang;JIANG Wei(School of Electronics and Information Engineering,Beijing Jiaotong University,Beijing 100044,China;State Key Laboratory of Rail Traffic Control and Safety,Beijing Jiaotong University,Beijing 100044,China;Beijing Engineering Research Center of EMC and GNSS Technology for Rail Transportation,Beijing 100044,China)
出处
《铁道学报》
EI
CAS
CSCD
北大核心
2022年第9期71-78,共8页
Journal of the China Railway Society
基金
国家重点研发计划(2018YFB1201500)
国家自然科学基金(61773049,61873023)。
关键词
列车运行控制系统
卫星导航
组合定位
最大相关熵
鲁棒估计
train operation control system
satellite navigation
integrated navigation
maximum correntropy criterion
robust estimation
