Due to large noise, severe multi-path error and frequent satellite of rise and fall caused by propagation obstacle and mirror material widespread in urban canyon with low-cost receiver, which was born with low-power design and restricted computing resource, the accuracy of position in urban canyon with normal RTK was heavily reduced with these large-biases observation and un-continuous tracking satellites, therefore it is not applicable to urban canyon without optimization. Therefore the C/N0-dependent exponential noise model for urban canyon was designed to balance practical noise and the noise model adopted in EKF filter, the accuracy of fixed-solution with this model was better than 0.5 m compared to former elevation-dependent model for majority of urban canyon but not applicable to special circumstance like deep urban canyon. Therefore, optimized empirical strategy for ambiguity resolution depending on posterior variance and number of satellite available was also proposed for special local circumstance like deep urban canyon, the results show that it has heavily reduced proportion of wrongly fixed solution, accuracy of fixed solution can be better than 0.10 m with greatly drop of wrong-fixed solution, and the accuracy of float-solution can be better than 0.898 m with good reliability and stability even if ambiguity resolution is failed. The proportion of correct-fixed solution in deep urban canyon was further reduced with ambiguity resolution AR-AIM method, which was proposed to migrate negative effect of frequent rise of satellite on ambiguity resolution.

中文翻译：

---

BDS/GPS RTK优化算法适用于城市峡谷低成本接收机

由于噪声大、多径误差严重、传播障碍物引起的卫星频繁起落，以及城市峡谷中普遍存在的镜面材料，低功耗设计和有限的计算资源，这些大偏差观测和非连续跟踪卫星在正常RTK的城市峡谷中的位置大大降低，因此不适用于未经优化的城市峡谷。因此设计了城市峡谷的 C/N0 相关指数噪声模型以平衡实际噪声和 EKF 滤波器采用的噪声模型，与以前的高程相关模型相比，该模型的固定解精度优于 0.5 m。大部分城市峡谷，但不适用于深城市峡谷等特殊情况。所以，针对城市深峡谷等特殊局部情况，还提出了基于后验方差和可用卫星数量的模糊度解决的优化经验策略，结果表明，它大大降低了错误固定解的比例，固定解的精度可以优于0.10 m，误定解大幅度下降，浮解精度优于0.898 m，即使模糊解法失败，可靠性和稳定性也很好。提出了模糊度分辨AR-AIM方法进一步降低了城市深部峡谷中正确固定解的比例，该方法旨在迁移卫星频繁上升对模糊度分辨率的负面影响。