Estimation of the Coefficients of the Error Formula for Long Baseline GPS Observations.

Abstract

In the case of GPS relative positioning, the errors in baseline vectors can be divided into constant and distance-dependent parts. Each component does not reflect a single error source, but a combination of error sources. This paper deals with the determination of the coefficients of constant and distance-dependent parts using observed data with the aid of Minimum Norm Quadratic Unbiased Estimation (MINQUE). These two coefficients are determined for a network of short baselines where the baseline lengths range from 35 to 130 km, and also for a network of long baselines ranging from 40 to 1, 600 km. For different networks, we get different values for the constant part, which depends on the type of observation, the types of receivers, the shape of the network and so on. We demonstrate how the distance-dependent part varies with the, baseline length. As in the case of the constant part, the estimated value of the coefficient of the distance-dependent part is different for different networks. However, it is shown that this coefficient is less than 1 ppm for the networks of longer and shorter baselines, and decreases with increasing baseline length.