Implementation of SLAM Algorithm Based on the Environment Features and Extended Kalman filter

SLAM algoritam je algoritam namijenjen istodobnoj lokalizaciji robota i izgradnji karte prostora bez potrebe za prethodnim prikupljanjem informacija o prostoru u kojem će se robot kretati. Pretpostavka kod EKF SLAM algoritma je mali broj značajki u prostoru zbog velike računske složenosti algoritma. Značajke su vrlo uočljivi elementi u prostoru pomoću kojih se mobilni robot može lako orijentirati i kasnije vršiti korekciju svog položaja. Pridruživanje podataka je termin koji predstavlja postupak preko kojeg mobilni robot može zaključiti da je značajka koju trenutno vidi, ista značajka koju je vidio prije, ali se sada zbog odometrijske greške nalazi na različitom položaju. Nakon toga se vrši korekcija robotovog položaja i značajki u prostoru.SLAM algorithm is an algorithm which purpose is to simoultaneously localize and map desired area without any information of environment in which the mobile robot is going to move. The assumption in EKF SLAM algorithm is small number of features in space because of its big calculation complexity. Features are very noticeable objects in space which the mobile robot can easily detect and later, using landmarks, make correction of its position. Data association is a statistical approach in which the mobile robot can make the conclusion that the landmark which it is currently seeing is the same landmark he saw before, but because of odometry error, it is currently positioned at different position than it was before. After that mobile robot can make correction of its pose