Abstract: To increase the accuracy of the large size mass property measurement system, a method of error compensation based on kinematics model is presented. The kinematics equation of the machine is established with D-H transform matrix and the relation between the measurement error and geometric parameters error is established. For some parameters which can be measured accurately, the error of them can be ignored, which simplify the calibration equation and calculating. To speed up the calibration process, optimal measure poses, which realize high positioning accuracy with small point number, are selected using particle swarm optimization algorithm(PSO). The relationship between pose of the calibration and fitness function is established. Experimental studies are performed to demonstrate the importance of PSO to the efficiency of calibration. Through the pose error compensation the max errors of the centroid, the moment of inertia and the product of inertia are reduced to approximately 10%, 50% and 20% respectively.

Key words: mass properties;kinematics;geometric parameters;calibration;particle swarm optimization algorithm;pose error;compensation