Comparison and optimization of algorithms for simultaneous localization and mapping on a mobile robot

master's thesis

Matic Rašl (Author), Aleš Holobar (Mentor), Eurico Farinha Pedrosa (Co-mentor)

Abstract

In this thesis, we compare and evaluate different SLAM solutions for a low-cost mobile robot. We present a simulator for the robot and use it to gather simulated data. Using this data, we then optimize the SLAM algorithms using an evolutionary algorithm. The optimized solutions are then validated and compared to default SLAM configurations. Up to 83 % reduction of error is achieved on validation data with multiple SLAM algorithms with improvements also visible on the real-world data.

Keywords

SLAM;optimization;mobile robot;evolutionary algorithm;simulation;

Data

Language:	English
Year of publishing:	2023
Typology:	2.09 - Master's Thesis
Organization:	UM FERI - Faculty of Electrical Engineering and Computer Science
Publisher:	[M. Rašl]
UDC:	004.021:004.94(043.2)
COBISS:	171288835
Views:	93
Downloads:	8
Average score:	0 (0 votes)
Metadata:

Other data

Secondary language:	Slovenian
Secondary title:	Primerjava in optimizacija algoritmov za sočasno lokalizacijo in kartiranje na mobilnem robotu
Secondary abstract:	V magistrskem delu primerjamo in ocenjujemo različne SLAM rešitve za nizkocenovnega mobilnega robota. Predstavimo simulator za robota in ga uporabimo za zbiranje simuliranih podatkov. Z uporabo teh podatkov nato optimiziramo algoritme SLAM z uporabo evolucijskega algoritma. Optimizirane rešitve se nato potrdijo in primerjajo s privzetimi konfiguracijami SLAM. Do 83 % zmanjšanje napake je doseženo pri validacijskih podatkih z več algoritmi SLAM z izboljšavami, vidnimi tudi pri podatkih iz resničnega sveta.
Secondary keywords:	algoritmi SLAM;optimizacija;mobilni robot;evolucijski algoritem;simulacija;magistrske naloge;
Type (COBISS):	Master's thesis/paper
Thesis comment:	Univ. v Mariboru, Fak. za elektrotehniko, računalništvo in informatiko, Računalništvo in informacijske tehnologije
Pages:	1 spletni vir (1 datoteka PDF (XIII, 69 f.))
ID:	19856817