Host: The Japan Society of Mechanical Engineers
Name : [in Japanese]
Date : June 28, 2023 - July 01, 2023
To achieve teaching-less disassembly by human operators and robots, we developed an automated sequence-generation method using only a 3D CAD model. First, to generate feasible and stable sequences, the proposed genetic algorithm generates a set of initial chromosomes based on contact and connection conditions rather than merely generating a set of random chromosomes. Second, to generate a sequence satisfying many preferred conditions under constraints, we defined constraints on feasibility and stability and designed objective functions regarding difficulty, efficiency, prioritization, and allocability. The ablation study results on sequence generation for a product (36 parts) demonstrated that the proposed method can generate feasible and stable sequences with a 100% success rate and sequences satisfying many preferred conditions needed for efficient disassembly by human operators and robots.
