REVISTA DYNA NEW TECHNOLOGIES

Key words:

optimización genética, robots móviles, calibración odométrica, genetic optimization, mobile robots, odometry calibration

Article type:

ARTICULO DE INVESTIGACION / RESEARCH ARTICLE

Section:

RESEARCH ARTICLES

ABSTRACT:
This paper presents a method for odometry calibration of tricycle topology mobile robots, which was focused on Automatic Guided Vehicle (AGV) . Odometry calibration for AGVs has been an open problem since the mobile robots inception. Basically, in this problem an AGV must have/know its position [x, y] and orientation [?] - broadly known as Pose- regarding to a starting point. Odometry calibration consists in determining the effective diameter of the AGV’s rear-wheels [k1, k2] and length of the axis between them [b]. These AGV parameters are used in a mathematical model to accurate calculate its position and orientation. The provided method get pulses from optical incremental-encoders (located at AGVs’ rear-wheels) in real-time, which are then used by a Genetic Algorithm to optimally determine the vector [k1, k2, b]. Obtained results suggested an absolute error of 1.5% (Max) for the pose calculation [x, y, ?] considering trajectories (straight and curved paths) up to 70 mts .

Key Words: genetic optimization, mobile robots, odometry calibration