Crankshaft-Bearing Assembly Based on Vision-Guided and Attractive Regions in the Environment

Abstract

Aiming at the problem of sensor-less high-precision manipulation, this paper designs and constructs a crankshaft-bearing assembly system based on vision-guided and attractive region in environment. A pre-analysis and evaluation of the attitude measurement accuracy method is proposed to describe the trusted region of the current assembly pose through vision. Then its high-dimensional attractive regions of environment constraints is constructed based on the mechanical constraints between the assembly object and the assembly position. Combine the complementary advantages of vision and mechanical information, a robot adjust its pose to a high precision manipulation pose. A platform of an industrial robot with a vision system is built and experiments on the crankshaft and bearing are successfully assembled. It is validated that using the proposed the vision-guided and attraction region in environment system, high precision assembly manipulation is realized.