广义阻尼LAMBDA方法及其在长基线实时定位中的应用

Generalized damped LAMBDA method and the application in long baseline real-time positioning

在长基线实时定位中,相邻观测值之间具有强相关性导致法方程出现病态,此时载波相位整周模糊度难以快速固定.针对这一问题,在传统阻尼最小二乘模糊度降相关平差(leastsquares ambiguity decorrelation adjustment,LAMBDA)方法基础上,提出了广义阻尼LAMBDA方法,并推导出载波相位模糊度的单历元解表达式.该方法利用了坐标和载波相位模糊度的先验信息,在改善法方程病态性的同时也提高了浮点模糊度的解算精度,有助于整周模糊度的快速固定.采用两条长基线(266 km和456 km)实测数据进行验证,改进后的方法相比于传统方法,载波相位模糊度固定率提高了28.8%;当放大因子c=0时,广义阻尼LAMBDA方法与传统阻尼LAMBDA方法等价.选择澳大利亚连续运行参考站系统(continuous operation reference system,CORS)网中四条长基线实时数据流进行测试,实验结果表明:在基线长度小于1 000 km的多系统实时定位中可以获得水平优于0.02 m,高程优于0.04 m定位精度.基于广义阻尼LAMBDA方法在长基线实时定位中具有一定的应用参考价值.

In long baseline real-time positioning,strong correlation between adjacent observations resulting in ill-conditioned normal equation which is adverse to fast integer solution of carrier phase ambiguity.In this contribution,a generalized damped least-squares ambiguity decorrelation adjustment(LAMBDA) method was be proposed based on traditional LAMBDA method and the solution expression for carrier phase ambiguity in single epoch also be derived.This method uses the prior information of coordinates and carrier phase ambiguity to improve the ill-condition of the normal equation and the solution accuracy of floating ambiguity,which is helpful for the fast integer ambiguity solution.Two long baselines(266 km and 456 km) measured data were used to verify.Compared with the traditional method,the ambiguity fixed rate of the improved method was increased by 28.8%.When the amplification factor c=0,the generalized damped LAMBDA method was equivalent to the traditional damped LAMBDA method.Four long baseline real-time data streams from the Australian continuous operation reference system(CORS)network were selected for testing.The experimental results show that multi-system real-time positioning accuracy was better than 2 cm in the horizontal direction and 4 cm in the elevation direction with a baseline length of less than 1 000 km.The generalized damped LAMBDA method has certain application reference value for long baseline real-time positioning.