基于UWB和惯性导航技术的无人车融合定位方法  

Fusion Positioning Method of Unmanned Vehicle Based on UWB and Inertial Navigation Technology

作　　者：商伟鹏 沈鸿平 刘斌辉 徐华伟[1,2] 陈超英 周茂杨 SHANG Weipeng;SHEN Hongping;LIU Binhui;XU Huawei;CHEN Chaoying;ZHOU Maoyang(CEPREI,Guangzhou 511370,China;Key Laboratory of MIIT for Intelligent Products Testing and Reliability,Guangzhou 511370,China)

机构地区：[1]工业和信息化部电子第五研究所,广东广州511370 [2]智能产品质量评价与可靠性保障技术工业和信息化部重点实验室,广东广州511370

出　　处：《电子质量》2023年第1期1-6,共6页Electronics Quality

基　　金：国家重点领域研发计划(2019YFB1804202);广州市基础研究计划(NO.202102021079);广东省重点领域研发计划(NO.2020B0404020001);工业和信息化部电子第五研究所2022年创客基金项目(2022CK07)资助。

摘　　要：针对基于超宽带(UWB)的无人车定位波动大和定位误差大的问题,提出了一种加权递推平均滤波和卡尔曼滤波相结合的定位算法。首先,建立了一套UWB基站定位系统,用于采集无人车UWB定位数据信息;然后,利用加权递推平均滤波算法对UWB定位数据信息处理,提升定位结果的稳定性;最后,结合无人车惯性导航信息对加权递推平均滤波算法处理后的结果进行卡尔曼滤波融合,得到最终的定位结果。经过无人车静止和移动场景的多组定位测试,结果表明,静止场景下加权递推平均滤波算法处理后定位均方根误差平均降低了13.3%,稳定性得到了一定的提升;移动场景下,采用加权递推平均滤波和卡尔曼滤波融合算法处理后,直线运动和圆周运动时的定位均方根误差平均分别降低了41.6%和38.4%,定位的稳定性和准确度得到有效的提升。由此证明了所提出的方法的有效性,具有一定的推广使用价值。Aiming at the problem of large positioning fluctuation and positioning error of unmanned vehicle based on UWB, a positioning algorithm based on weighted recursive average filtering and Kalman filtering is proposed. Firstly,a set of UWB base station positioning system is established to collect the UWB positioning data of unmanned vehicle.Then, the weighted recursive average filtering algorithm is used to process the UWB location data information, which is helpful to improve the stability of the location results. Finally, combined with the inertial navigation information of unmanned vehicle, the results of weighted recursive average filtering algorithm is fused by Kalman filtering, and the final positioning results are obtained. The results of multiple localization tests in static and moving scenes of unmanned vehicles show that the root mean square error of localization is reduced by 13.3% on average after processing by weighted recursive average filtering algorithm in static scenes, and the stability is improved to a certain extent. In the moving scene, after the fusion processing of weighted recursive average filtering and Kalman filtering algorithm, the average root mean square error of the location in linear motion and circular motion is reduced by 41.6% and 38.4% respectively and the stability and accuracy of positioning are effectively improved. And it is proved that the proposed method is effective and has a certain value of popuarization and application.

关 键 词：无人车 超宽带定位 加权递推平均滤波 卡尔曼滤波融合 数据融合 

分 类 号：TN925[电子电信—通信与信息系统] TN713[电子电信—信息与通信工程]