【摘要】 随着智能设备的普及,使用手机为汽车进行导航越来越普遍。目前,手机内置的惯性测量单元成本低廉、导航精度低,且常使用松组合,而应用GPS/SINS紧组合导航技术可提高导航的精度及可用性。通过将基于Newton迭代的梯度下降算法与滚动时域估计相结合,手机内置GPS与陀螺加速度计构成紧组合导航系统,可提高在复杂环境中的导航精度。传统Kalman滤波方法未能充分考虑到紧组合导航系统的非线性特性,而滚动时域估计通过参考多时刻导航信息,可更有效地排除非线性因素对系统的干扰,并通过实验验证了此方法。实验结果表明,相比采用扩展Kalman滤波,基于Newton迭代时域估计算法的紧组合导航系统的导航位置精度及速度精度均获得了较大提高。更多还原

【Abstract】 With the popularization of smart devices,there is much more widespread use of cell phone navigation in vehicles.Nowadays,inertial measurement units in mobile phone are low-cost with low precision,and a loosely-coupled scheme is usually applied.Thus,tightly-coupled navigation composed by phones ’ internal GPS,gyroscopes and accelerometers would contribute to improving navigation accuracy and availability.An algorithm combining Newton method and moving horizon method is introduced to aid tightly-coupled navigation system,to improve navigation accuracy under complicated circumstances.The traditional extended Kalman filters can not fully take nonlinear characteristic of tightly-coupled navigation system into account,thus moving horizon method is introduced by referring to multi-time navigation information to eliminate interface caused by nonlinear factor.The new algorithm is tested through actual experiment.It turns out that the Newton method based moving horizon estimation algorithm of tightly-coupled navigation system achieves substantial improvement in positioning accuracy and velocity accuracy than extended Kalman filter.更多还原