摘要
在接收信号强度指示(RSSI)指纹定位与行人航位推算(PDR)组合室内定位模型中,针对噪声误差和模型误差对融合定位的影响,提出了一种基于混合H2/H∞滤波的融合定位方法。由于室内信号存在多径传播、衰减、非视距(NLOS)的影响以及传感器精度问题,系统或观测噪声往往不符合卡尔曼滤波(KF)的假设前提,造成滤波稳健性差。H∞滤波不对噪声特性和模型误差做出假设,具有一定抗干扰能力,但由于结果保守,精度略低。混合H2/H∞滤波同时兼顾了二者的特性,在精度和鲁棒性提供一种均衡。通过仿真实验对比这3种滤波的融合定位性能。实验结果表明:混合H2/H∞滤波有效提高了指纹/PDR组合室内定位系统的精度和鲁棒性。
In combined indoor localization model of received signal strength indication(RSSI)fingerprint localization and pedestrian dead reckoning(PDR),a fusion localization method based on hybrid H2/H∞filtering is proposed,aiming at the influence of noise error and model error on fusion localization.Due to multi-path propagation,attenuation,non-line-of-sight(NLOS)influence of indoor signals and sensor precision problems,the system or observation noise often does not conform to the prerequisite of assumptions of Kalman filtering,resulting in poor filtering robustness.H∞filtering does not make assumptions about noise characteristics and model errors,and has a certain anti-interference ability,but due to result is conservative,the precision is slightly lower.Mixed H2/H∞ filtering takes into account the characteristics of both,and provides an equilibrium in precision and robustness.Fusion positioning performance of three filtering is compared by simulation experiment.Experimental results show that the mixed H2/H∞filtering effectively improves the precision and robustness of the fingerprint/PDR integrated indoor positioning system.
作者
张呈
张大龙
刘成明
乔寅嵩
孙顶
ZHANG Cheng;ZHANG Daong;LIU Chengming;QIAO Yinsong;SUN Ding(School of Cyber Science and Engineering,Zhengzhou University,Zhengzhou 450002,China)
出处
《传感器与微系统》
CSCD
北大核心
2024年第1期139-144,共6页
Transducer and Microsystem Technologies
基金
国家重点研发计划资助项目(2020YFB1712401)。
关键词
室内定位
多源信息融合
混合H2/H∞滤波
指纹定位
行人航位推算
indoor positioning
multi-source information fusion
mixed H2/H∞filtering
fingerprint positioning
pedestrian dead reckoning(PDR)
