Analyses on on-line evolutionary optimization performance for pose tracking while eye-vergence visual servoing

In this research, Genetic algorithm (GA) is used as pose-tracking method “1-step GA,” to solve on-line optimization problem for 3-D visual servoing. A correlation function between the target object projected in camera flame and model defined in the control computer is used for a fitness function to be optimized by the 1-step GA. The optimization process for real-time object tracking has been examined on a view point of realizing real-time pose estimation utilizing eye-vergence function. We have confirmed that the 1-step GA optimization method together with eye-vergence correlation fitness function worked cooperatively and how the eye-vergence helped real-time optimization processes in time-domain during visual servoing.