多传感器组合是提高导航系统定位精度和增强系统容错性的有效手段。在分析惯性、星光、卫星导航工作特性的基础上,提出了基于集中滤波器结构的捷联惯性/星光/卫星信息融合导航方案。针对多传感器信息融合中非等间隔量测特性,设计了时间更新和量测更新分离的异步集中卡尔曼滤波算法,设计了基于外推法的卫星信息补偿算法,有效解决了多传感器非等间隔信息融合问题;针对垂直机动研究了惯性/星光姿态组合模型。仿真结果表明,算法可以有效实现对捷联惯导、星光、卫星导航信息的融合,组合精度提高1倍,具有重要的实际应用价值。
Multi-sensor combination is an effective means to improve the accuracy and fault tolerance of navigation system. Based on the analysis of SINS, STAR and GPS, a SINS/STAR/GPS information fusion navigation system is presented. To solve the incoordinate interval problem of multi-sensors, an asynchronous centralized Kalman filter is designed, and the filtering period is divided into time update period and measurement update period. An extrapolation method is designed to deal with GPS information. Moreover, a new model is designed to solve the problem of attitude combination in the process of vertical mobility. Test results indicate that the filter accuracy is improved by 100% with the Kalman filter, and the method has important value in engineering application.