Application of Genetic Algorithm with Optimization of GNSS Satellite Combination in Kinematic Positioning Mode: Case Study using GPS, GLONASS and Beidou Data.

Abstract

Kinematic GNSS positioning mode has been widely used in many applications. There are various techniques to obtain high-precision kinematic positioning results. One such technique is relative positioning based on differential carrier phase-based positioning. This technique may produce high accuracy positioning solutions in good, obstruction free signal environments. However, in less favorable observing environments (i.e., urban areas with tall buildings or tree canopy), the remaining errors can cause bad or unreliable positioning results. The quality of the solution mainly depends on the ability to resolve the ambiguities to their correct integer values. The process is known as Ambiguity Resolution (AR). AR success relies on the removal of unreliable observations data before data processing among others. Generally, the removal is carried out manually and may be considered as a trial-and-error procedure. The user has to process over and over until he gets a satisfactory result. Obviously, the manual removal is time consuming. In addition, it requires skills of an experienced user. To avoid such challenges, an automatic procedure is recommended. This paper introduces an optimization procedure based on the genetic algorithm (GA). GA is a global optimum search algorithm based on natural evolution that has been extensively used in various fields of study. GA allows optimally selecting the best GNSS satellite combination and improving the percentage of the ambiguity-fixed solutions in kinematic positioning mode. The numerical results demonstrate the improvements in ambiguity resolution rate when GA is applied comparing to the conventional method. Overall, the proposed GA-based method can improve the ambiguity resolution rate at 48.85 percent, 50.08 percent and 36.79% for 3.8 km, 6.7 km and 23.2 km baseline lengths respectively.