基于卫星载噪比定权的手机RTK/PDR融合定位研究

针对智能手机GNSS定位芯片在遮挡环境下定位精度下降的问题,本文采用RTK/PDR融合方法解算位置结果,考虑到GNSS信号易受环境影响,本文提出根据平均卫星载噪比参数将当前GNSS观测环境分为优、中、差三种情况,并根据判断出的不同观测条件,设置不同的膨胀系数放大Kalman融合滤波中量测噪声方差矩阵R,实现在遮挡环境中对GNSS位置信息的降权处理,以获得更优的融合结果.实验测试显示:在开阔环境中,两部测试手机RTK与RTK+PDR定位精度基本一致;在轻微遮挡环境中,两部测试手机RTK+PDR融合定位结果较RTK单独定位结果平面提升效果分别为6.24%和3.97%,高程提升效果分别为20.44%和32.56%;在严重遮挡环境中,两部测试手机RTK+PDR融合定位结果较RTK单独定位结果平面提升效果分别为47.65%和42.90%,高程提升效果分别为39.78%和67.98%.实验结果表明,本文提出的依据平均卫星载噪比动态设置GNSS信息权重的RTK+PDR融合定位方法,能够部分弥补手机RTK在遮挡地区定位精度下降的缺点.

Smart phone RTK/PDR fusion positioning based on satellite carrier noise ratio weighting

Aiming at the low positioning accuracy of the GNSS positioning chip of the smart phone in occluded environment,this paper uses the RTK/PDR fusion method to solve the position results.Considering that the GNSS signal is susceptible to environment noise,different expansion coefficients are set to amplify the measurement noise variance matrix R in Kalman fusion filtering to realize the weight reduction of GNSS position information in occluded environment,so as to obtain better fusion results.The experimental test shows that in open environment,the positioning accuracies were identical between RTK and RTK+PDR for both two test mobile phones.While in the slightly occluded environment,compared with the RTK single positioning,the RTK+PDR fusion positioning improved the plane lifting effect of the two test mobile phones by 6.24% and 3.97%,and improved the elevation lifting effect by 20.44% and 32.56%.In the severely occluded environment,the RTK+PDR fusion positioning improved the plane lifting effect of the two test mobile phones by 47.65% and 42.90% compared with the RTK single positioning,and improved the elevation lifting effect by 39.78% and 67.98%.The experimental results show that the RTK + PDR fusion positioning method,which dynamically sets the GNSS information weight according to the average satellite carrier to noise ratio,can partially compensate for the reduction of positioning accuracy caused by environment occlusion for mobile phones using RTK.