Simulace lokalizačního algoritmu nemocničního robotu

Abstract

This diploma thesis concerns itself with localization methods of mobile medical robots using laser snapshot, simultaneous localization and mapping (SLAM). At the beginning of this paper various existing methods for snapshot alignment are introduced and categorized. The goal of this diploma thesis is to propose a new method for medical robot localization using evolution algorithm in connection with cross-correlation. Differential evolution was selected as an evolution algorithm. This method is based on creation of new populations and evolution of new generations of differential evolution for the purpose of snapshot alignment. Plausibility of the mutation candidate from a given population is evaluated using cross-corelation, correlation coefficient for three-dimensional correlation. Subsequently, a number of experimental measurement is conducted to show method robustness and computational complexity compared to an existing (benchmark) algorithm.