Optimal Grasping Pose Selection Method for Dual-arm Robot Based on Improved Genetic Algorithm

Abstract

Dual-arm robot has been widely used in industry and service trades. It increases the degree of freedom within the workspace, while leading to more complex task planning problems. When setting goals for the dual-arm, it is a key issue to consider the impact of the setting of the goals on the complexity of the task. In this paper, an optimal grasping pose selection method has been proposed in order to select the optimal grasping pose of the dual-arm robot. This method uses an improved genetic algorithm. Facing the task of multi-objective optimization, the fitness function and gene reservation strategy can be adjusted automatically according to the iterative depth. Thereby, the coordinates and grasping pose of the arms on the object are obtained. The simulation experiment of dual-arm robot grasping slender objects was carried out. The results show that it has a better performance in symmetry of grasping points, position variation and synchronization of dual-arm robot.