The optimization of fitting a mobile robot’s position and pose to its measured coordinates’ data by single vision

This paper is about a kind of mobile robot used to polish free form surface of large metal model above and a method is presented for fitting the robotpsilas configuration to the measured data obtained by machine vision system. In the close loop of control system of the robot, a single camera machine vision system is employed to measure robotpsilas position and pose. Due to 2D state data can only be obtained by signal camera, the real 3D parameters of robot must be calculated out when it is moving on uneven surface. But errors usually exist in the calculated data. Thus to obtain more accurate data about robotpsilas state, it is needed to fit the measured data group with the robotpsilas configuration. Base on genetic algorithm (GA), an optimal fitting method is proposed and better result is got.