Quality Control with Reliability for Large GPS Networks

Abstract

The Orange County GPS survey comprises more than 7,000 vectors linking more than 2,000 stations. The observations were made by private firms and the County of Orange Survey Division staff. Several functions are identified and used to analyze this large network. Weak portions in the network are identified by outliers in these functions. The variance‐covariance matrix of the vector observations is the “determining force” for shaping the functions. Redundancy numbers and internal reliability plots appear to be especially useful in identifying weak portions of the network (which may result from automated deweighting for blunder detection). Small residuals are not the only criteria for judging global positioning system (GPS) networks. Equally important is the capability of the network to isolate blunders. Several graphs are presented in order to shed light on the stochastic and geometric properties of GPS vector networks.