Comparison of mobile robot positioning techniques

Abstract

The article compares the accuracy of mobile robot positioning by the technique based on genetic algorithms, which are related to artificial intelligence, and by the trilateration technique. The authors consider the application of appropriate terminology borrowed from genetics and data processing algorithms for this technical problem. When using the genetic algorithm, the coordinates of the robot are found using angular methods or rigid logic methods, which are not particularly effective because of the large amount of data that is not needed for positioning, so there is a need to select the most likely indicators to find the best route to the target. The genetic algorithm used in this study first selects the data by a certain criterion to enter the first population, and then the data falls into the beginning of the genetic algorithm. Each individual has chromosomes that represent a sequence of data, i.e., genes. After a chromosome is coded, the following genetic operations are performed: crossing over and mutation. These operations occur cyclically until a population with high fitness is found. The solution is a sequence of selected coordinates, from which a system is constructed to determine the optimal route to the destination. The robot navigation techniques are compared in terms of coordinate positioning accuracy. Calculation results on dispersion and absolute positioning error show that the positioning using genetic algorithm gives less error than the one using trilateration method. The genetic algorithm allows finding the optimal solution of the positioning problem while reducing a significant influence of the measurement error of sensors and other measuring devices on the result.