Combining Vision Sensing with Knowledge Database for Environmental Modeling in Dual-arm Robot Assembly Task

In this paper, focusing on the vision-based dual-arm collaborative assembly, a knowledge-based representation model is proposed. Different from the traditional assembly behavior in which the assembly parameters are obtained by fixing the assembly object at a specified position in a specified posture, the proposed model uses the prior knowledge bound to the model to generate assembly parameters through point cloud matching and calculation. To a certain extent, the dependence and restriction on the fixing device during the assembly process are eliminated. And the proposed is verified on ABB YuMi. The simulations prove that the proposed approach can achieve a good performance for assembly tasks.