Second order design of geodetic network using different TK structured criterion matrices

Geodetic network design for numerous practical applications (surveying, staking out of characteristic points of an object in construction, control of geometry of a built-up object, monitoring, etc) represents a problem which requires implementation of well-known optimization methods. Optimization aim is establishing as much accurate and reliable geodetic network as possible with as less costs as possible. Geodetic network optimization is classified into different orders but in this paper only the second order design which deals with the determination of the weights of network measurements is presented. Within the second order design it is necessary to form a criterion matrix which substitutes real covariance matrix and to define observation scheme in network. The criterion matrix is formed by using one of the correlation functions whose arguments are distances between geodetic network points. This paper presents comparative analysis of the results of the second order design of trilateration geodetic network. These results were obtained by using TK-structured criterion matrices formed by using Gaussian and Baarda’s correlation functions of point coordinates. As it was expected, Gaussian correlation function is the one that proved as a better choice because it generally adapts better to the geodetic network requirements than Baarda’s correlation function: the results obtained by using Gaussian correlation function tolerate less accurate distance measurements.