| 摘要: |
| 针对容积卡尔曼滤波在多源融合定位中存在跟踪能力不强和自适应能力差的问题,在传
统容积卡尔曼滤波的基础上,提出了改进自适应抗差容积卡尔曼滤波算法。建立了基于新息的自
适应判决准则与修正方法,使得滤波算法能够及时跟踪目标真实状态;引入抗差因子调节观测协
方差矩阵,以减小观测值异常问题对滤波精度的影响;采用奇异值分解代替容积卡尔曼中的Cholesky
分解,提高数值计算的稳定性。超宽带/惯性导航联合定位实验结果表明,与扩展卡尔曼滤波
和容积卡尔曼滤波相比,改进的自适应抗差容积卡尔曼滤波定位精度更高,数值稳定性更好,增强
了定位系统在粗差干扰下的鲁棒性。 |
| 关键词: 容积卡尔曼滤波 奇异值分解 自适应修正 抗差因子 室内定位 |
| DOI： |
|
| 基金项目:国家自然科学基金(61773330);国家重点研发计划(2020YFA0713501);湖南省教育厅科学研究项目
(18C0126);湖
南省自然科学基金(2021JJ50126);湖南省教育厅重点项目(21A0083) |
|
| Improved Adaptive Robust Cubature Kalman Filter forMulti-source Indoor Positioning |
| LIPeng,RONGDong-cheng,XIANGYu-xiang,LINGZhi-chen,XIAJun |
| (CollegeofAutomationandElectronicInformation,XiangtanUniversity,Xiangtan,Hunan411100,China) |
| Abstract: |
| CubatureKalmanfilterhastheproblemsofweaktrackingabilityandpooradaptiveabilityin
multi-sourcefusionpositioning.BasedonthetraditionalcubatureKalmanfilter,animprovedadaptiverobustcubatureKalmanfilteralgorithmisproposedinthispaper.
Theadaptivedecisioncriterionandcorrectionmethodbasedoninnovationareestablished,
sothatthefilteringalgorithmcantracktherealstate
ofthetargetintime.Therobustfactorisintroducedtoadjusttheobservationcovariancematrixtoreduce
theinfluenceofobservationanomalyonthefilteringaccuracy. Singularvaluedecompositionisused
insteadofCholeskydecompositionincubatureKalmanfiltertoimprovethestabilityofnumericalcalculation.
TheexperimentalresultsofUWB/inertialnavigationjointpositioningshowthatcomparedwithextendedKalmanfilterandcubatureKalmanfilter,
theimprovedadaptiverobustcubatureKalmanfilterhas
higherpositioningaccuracyandbetternumericalstability,andenhancestherobustnessofthepositioning
systemundergrosserrorinterference. |
| Key words: CollegeofAutomationandElectronicInformation,XiangtanUniversity,Xiangtan,Hunan411100,China |
