Virtual Interface Technology in the Process of Simulation of Complex Functional Modules of Control Systems for Industrial Robots and Multi-Axis Mechatronic Systems

Abstract

Purpose of research. Development of a tool for debugging intelligent control system algorithms, including the development of a vision system and planning a software trajectory for an industrial robot.Methods. To achieve this goal, a review of existing simulation tools was carried out. A protocol of contactless humanrobot interaction is presented. An algorithm for the recognition of gesture commands based on the difference of three-dimensional binary microblocks and the construction of the skeleton of the human body has been developed. An example of using the ROBOGuid software tool for imitating the motion of an industrial robot in the development and debugging of its own control methods focused on real objects is presented.Results. The use of digital twins of technological equipment to simulate and display real technological processes in a virtual environment, in the context of the formation of a new concept of Industry 4.0 and the sixth technological order, allows improving the main and auxiliary production processes, as well as analyzing, researching and evaluating the economic efficiency of new technological and technical solutions. Simulation allows the development of ergonomic ways of human interaction with mechatronic objects. The solution proposed in the work was tested on the example of working out a complex spatial contour that simulates the milling of a part. Experimental studies of the gesture command recognition algorithm proposed in the work were carried out on the publicly available UCF101 dataset, the results are compared with known approaches to recognizing human actions.Conclusion. The developed interface module was used on the example of working out a complex spatial contour that simulates the milling of a part, and the method of a contactless robot control system has shown its effectiveness and the need to develop this direction.

Action recognition algorithm from multimodal images for human-robot collaboration systems

“…The algorithm is a four-step procedure: [1] Fusion information from multimodal sensors based on a parameterized model. [2] Image preprocessing by the Gabor Quaternionic Fourier transform.…”

“…The developed software solution is implemented in the Python programming language. To develop algorithms for recognition of gesture commands and study their effectiveness, simulation modeling was carried out in the ROBOGuide environment [1] by the Fanuc company for contactless control of robot movements [2]. For field experiments, the FANUC M20iD robot presented in the laboratory of collaborative robotics of MSTU "STANKIN" was used.…”