A Method of Virtual Assembly Concentric Ball Scanning Path Planning Method Based on Depth Learning

Abstract

In virtual assembly, it is important to select and determine the best path of assembly. This paper proposes a concentric sphere scanning path planning method based on depth learning, which takes the datum part as the functional center and the radius difference between the two concentric spheres as search step length. Assembly constraints and five-first guideline are constructed mainly with the special sequence and functional sequence. Topological relation tree and relations are used. All the assembling elements are vectorized and digitalized. The spindle assembly is taken as an example for verification. Results show that the method proposed in the current study is of effectiveness.