Robot kinematic calibration using Genetic Algorithms
Position and orientation accuracy of the end effector is affected by
the precision of kinematic parameters of the robot manipulator. Thus
good precision requires good knowledge of robot physical parameter
values. However this condition can be difficult to meet in practice.
Hence calibration techniques can be devised in order to improve the
robot accuracy through estimation of these particular parameters. In
this paper, the Genetic Algorithm is used to calibrate the robot
kinematic accuracy. A kinematic model is formulated and conducted as an
optimization problem for ABB robot manipulators. The objective is to
analyze and evaluate the performance of the GA in optimizing such robot
kinematic accuracy. In this algorithm, small changes in the kinematic
parameters values represent the parent and offspring population and the
end-effector error represents the fitness functions. A numerical
example has been used to demonstrate the convergence and effectiveness
of the given model.
I am sorry to reply the questions so later!
Our method is to calibrate robot kinematic parameters using GA approach. We consider the calibration process as a search for optimized solution. This approach may be run once when the robot manipulator leaves out the factory.
Who's online
There are currently 0 users and 1125 guests online.
Good to see a paper that tries different GA approaches. Which calibration is this method designed to replace? Just once when it leaves the factory, or whenever the robot hits something and needs to be re-aligned?