Assembly skill direct generation method based on virtual assembly environment

Abstract

The invention discloses an assembly skill direct generation method based on a virtual assembly environment, and the method comprises the steps: an assembly demonstrator controls a mechanical arm to work in a first-person operation mode, carries out the assembly operation of an industrial product in a virtual environment through a tool and process equipment which accord with a real assembly environment; for entities such as end effectors, parts, tools, process equipment, parts and assemblies involved in the assembling process and operations such as alignment, approaching, grabbing, inserting holes and screwing between the entities, a dynamic operation knowledge graph is directly generated according to defined entities and operations, and a primary framework of the operation knowledge graph is established; knowledge updating can be continuously carried out by operating the knowledge graph, and new entities and actions are added into the ontology library. The knowledge graph can be used as a standard format of operation skills and is used for operation skill transmission among different agents, and the method can also be expanded to other operation skill learning fields.

Description

technical field [0001] The invention belongs to the field of learning operation skills in artificial intelligence, and relates to a method for directly generating assembly skills based on a virtual assembly environment, which is used for an intelligent body to teach assembly operation skills based on human operation demonstration. Background technique [0002] Robot manipulation skill refers to the ability of a robot to operate a specific object in the environment within a limited time based on its own sensing, perception, decision-making, planning and control capabilities, so that the object can reach the target state from the initial state. Skills are ubiquitous in life and production, such as depalletizing and palletizing in logistics, assembly of parts on the production line, serving tea and water in restaurants, and playing table tennis in sports. As artificial intelligence technology develops, the need to teach robots complex skills increases. If robots can replace or...

AI Technical Summary

Problems solved by technology

Although the use of end-to-end or existing self-learning methods has good research potential, due to the limitations of the method, it cannot be promoted and applied in the industrial field; [IKEA Furniture Assembly Environment for Long-Horizon Complex Manipulation Tasks] Youngwoon Lee et al. Constructed a virtual assembly environment for IKEA furniture. Although it can complete many assembly tasks, the assembly target is limited to furniture, but knowledge extraction and other operations are required to generate knowledge, and operational skills cannot be transferred; [RLBench: The Robot Learning Benchmark&Learning Environment] Stephen James proposed a robotic reinforcement learning virtual environment RLBench capable of editing learning tasks. Although the method is somewhat forward-looking, it does not utilize the direct manipulation data generated in the virtual environment.

Images