Virtual operation demonstration system for man and robot skill transmission

Abstract

The invention discloses a virtual operation demonstration system for human and robot skill transmission. The system is mainly composed of three parts of virtual scene modeling, job information captureand knowledge mining analysis. A home work scene is built based on the virtual reality technology, whole-body motion capture and environment global information perception of a demonstrator are achieved, a behavior hierarchical multi-weight segmentation recognition algorithm combining motion features and operation object semantics is designed to obtain task behavior high-level semantic knowledge,and a task behavior state machine is generated. The virtual operation demonstration system provided by the invention has the advantages of high efficiency, low cost and the like, not only can providerich, perfect and accurate operation demonstration data for humanoid behavior learning of the service robot, but also can mine task knowledge such as an operation sequence, an operation target and thelike; the system has important significance in promoting construction and maintenance of a robot task knowledge base and improving applicability and flexibility of the robot.

Description

technical field [0001] The invention relates to the construction field of a virtual job demonstration system, in particular to a virtual job demonstration system for man-machine skill teaching. Background technique [0002] With the aging of the society and the increasing shortage of labor force, robots will not only play a more important role in the traditional industrial field, but also gradually enter the daily life of human beings, adapt to the living environment of human beings, and become an indispensable close friend of human beings and helpers. Different from industrial robots, service robots mostly work in a dynamic and unstructured environment. The tedious daily work is difficult to be strictly regulated by the program. The task objectives and indicators are more diverse, the operation flexibility is higher, and the task process is more complex and changeable. [0003] Regarding daily work, humans have accumulated a lot of knowledge and experience. If robots can u...

AI Technical Summary

Problems solved by technology

[0004] However, obtaining human behavior demonstration data requires building different demonstration scenarios according to tasks, and not only does the job scene have diverse object requirements, but the work of adding labels to all scene objects is also complicated and cumbersome, which greatly increases the workload; at the same time, in order to capture Complete, accurate and diverse job demonstration data such as "human body movement information", "three-element interaction information of people-object-environment" and "surrounding scene dynamic information" need to be equipped with a perception system covering the whole scene, but the realization of full-scene Perceptual coverage requires a variety of high-precision sensors to cooperate with each other, such as visual recognition systems and high-precision optical motion capture systems, etc., which will greatly increase the difficulty of building demonstration systems and application costs, making it impossible to popularize and limit human-like behavior of robots Research Advances in Learning Technology

In addition, the existing job demonstration system pays more attention to the motion information of the job, but ignores the semantic information of the behavior and the scene, lacks semantic extraction and knowledge reasoning capabilities, and is difficult to extract behavioral knowledge for long-term complex job demonstrations. Acquiring high-level task abstract knowledge

Images