College of Automation, Harbin Engineering University, Harbin 150001, China.
Department of Geomatics, University of Calgary, Calgary, AB T2N 1N4, Canada.
Sensors (Basel). 2018 Dec 6;18(12):4305. doi: 10.3390/s18124305.
This paper implements and analyzes a tightly coupled single-frequency global navigation satellite system precise point positioning/inertial navigation system (GNSS PPP/INS) with insufficient satellites for land vehicle navigation using a low-cost GNSS receiver and a microelectromechanical system (MEMS)-based inertial measurement unit (IMU). For land vehicle navigation, it is inevitable to encounter the situation where insufficient satellites can be observed. Therefore, it is necessary to analyze the performance of tightly coupled integration in a GNSS-challenging environment. In addition, it is also of importance to investigate the least number of satellites adopted to improve the performance, compared with no satellites used. In this paper, tightly coupled integration using low-cost sensors with insufficient satellites was conducted, which provided a clear view of the improvement of the solution with insufficient satellites compared to no GNSS measurements at all. Specifically, in this paper single-frequency PPP was implemented to achieve the best performance, with one single-frequency receiver. The INS mechanization was conducted in a local-level frame (LLF). An extended Kalman filter was applied to fuse the two different types of measurements. To be more specific, in PPP processing, the atmosphere errors are corrected using a Saastamoinen model and the Center for Orbit Determination in Europe (CODE) global ionosphere map (GIM) product. The residuals of atmosphere errors are not estimated to accelerate the ambiguity convergence. For INS error mitigation, velocity constraints for land vehicle navigation are adopted to limit the quick drift of a MEMS-based IMU. Field tests with simulated partial and full GNSS outages were conducted to show the performance of tightly coupled GNSS PPP/INS with insufficient satellites: The results were classified as long-term (several minutes) and short-term (less than 1 min). The results showed that generally, with GNSS measurements applied, although the number of satellites was not enough, the solution still could be improved, especially with more than three satellites observed. With three GPS satellites used, the horizontal drift could be reduced to a few meters after several minutes. The 3D position error could be limited within 10 m in one minute when three GPS satellites were applied. In addition, a field test in an urban area where insufficient satellites were observed from time to time was also conducted to show the limited solution drift.
本文使用低成本 GNSS 接收器和基于微机电系统 (MEMS) 的惯性测量单元 (IMU)，实现并分析了一种在陆地车辆导航中使用不足卫星的紧耦合单频全球导航卫星系统精密单点定位/惯性导航系统 (GNSS PPP/INS)。对于陆地车辆导航，不可避免地会遇到可观测卫星不足的情况。因此，有必要分析 GNSS 挑战性环境下紧耦合集成的性能。此外，研究采用最少数量的卫星来提高性能，与不使用卫星相比，也很重要。本文使用低成本传感器进行紧耦合集成，在没有任何 GNSS 测量的情况下，清楚地观察到了采用不足卫星的解决方案的改进情况。具体来说，本文使用单频接收机实现单频 PPP，以获得最佳性能。INS 机械化为局部水平框架 (LLF) 进行。应用扩展卡尔曼滤波器融合两种不同类型的测量值。更具体地说，在 PPP 处理中，使用 Saastamoinen 模型和欧洲轨道确定中心 (CODE) 全球电离层图 (GIM) 产品校正大气误差。不估计大气误差残差以加速模糊度收敛。为了减轻 INS 误差，采用陆地车辆导航的速度约束来限制基于 MEMS 的 IMU 的快速漂移。进行了具有模拟部分和完全 GNSS 中断的现场测试，以展示具有不足卫星的紧耦合 GNSS PPP/INS 的性能：结果分为长期 (几分钟) 和短期 (不到 1 分钟)。结果表明，一般来说，应用 GNSS 测量时，尽管卫星数量不足，但仍可以改善解决方案，尤其是观测到的卫星数量多于三颗。使用三颗 GPS 卫星，在几分钟后，水平漂移可以减少到几米。当应用三颗 GPS 卫星时，在一分钟内可以将三维位置误差限制在 10 米以内。此外，还进行了一次在城市地区进行的现场测试，该地区不时观察到卫星不足的情况，以显示有限的解决方案漂移。
