基于约束优化的多传感器车载定位方法  

Multi-sensor Vehicle Positioning Algorithm Based on Constraint Optimization

作　　者：张子腾 盛传贞[1,2] 蔚保国 王星星[1,2] 易卿武 ZHANG Ziteng;SHENG Chuanzhen;YU Baoguo;WANG Xingxing;YI Qingwu(The 54th Research Institute of CETC,Shijiazhuang 050081,China;State Key Laboratory of Satellite Navigation System and Equipment Technology,Shijiazhuang 050081,China)

机构地区：[1]中国电子科技集团公司第五十四研究所,河北石家庄050081 [2]卫星导航系统与装备技术国家重点实验室,河北石家庄050081

出　　处：《无线电工程》2022年第10期1781-1787,共7页Radio Engineering

基　　金：国家重点研发计划(2019YFC1511504)。

摘　　要：基于多传感器融合定位平台,采用误差互补的卡尔曼滤波或误差优化方法是解决室内外无缝定位的主流手段。但是,在缺乏先验地图的城市峡谷和山区等遮蔽、半遮蔽环境,应急救援车载定位系统仍然面临频繁失效问题。基于此,提出了一种基于约束优化的多传感器车载室内外定位方法,设计联合GNSS/IMU/激光雷达的多源融合定位系统,在状态更新模型的基础上,分析了车辆连续起伏过程受力方程约束低成本MEMS高程震荡,设计引入RTK解析状态监测评估约束GNSS可用性,提出联合各传感器误差约束构建集束调整优化方程;分别进行室内、城市和山区场景的实时车载定位测试。实验结果表明,该方法可实现亚米级室内定位和车道级室外定位。With the help of multi-sensor platform,it is the mainstream method to solve indoor and outdoor seamless positioning based on error complementary Kalman filtering or error optimization.However,emergency rescue vehicle positioning systems that lack priori map are faced with frequent failures in sheltered or semi-covered environments such as urban canyons and mountain areas.Based on the above issue,a constraint optimization algorithm for multi-sensor indoor and outdoor vehicle positioning,and a multi-sensor fusion positioning system by GNSS/IMU/Lidar is designed.Based on the state update model,a force equation is analyzed when the vehicle is traveling on an undulating road to constrain low-cost MEMS elevation oscillations,RTK analysis state monitoring and evaluation is introduced to constrain GNSS availability,and a bundle adjustment optimization equation is constructed via the error constraint of each sensor.Finally,real-time vehicle positioning tests in indoor,urban and mountain scenes are conducted,and the experiments show the algorithm can achieve sub-meter indoor positioning and lane-level outdoor positioning.

关 键 词：多传感器 车载定位 监测评估 集束调整 室内外 应急救援 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]