Aug 5, 2019 - FEFU News
FEFU scientists taught industrial robots to make decisions independently in a varying environment

Scientists at the School of Engineering of Far Eastern Federal University (FEFU), the Institute of Marine Technology Problems and the Institute of Automation and Control Processes of the Far Eastern Branch of the Russian Academy of Sciences (FEB RAS) have developed a software that enables industrial robots with artificial intelligence, including a technical vision system, to independently in real-time form and adjust the trajectories of the movements of their tools, while maintaining a given accuracy of movements.

The report on this topic was recognized as the best in its section at the ICCAD’19 conference, which was held in Grenoble, France from the 2nd to 4th of July.

FEFU scientists have created and put into practice a new principle of intelligent control using industrial robots - control of program signals. This means robots themselves can determine and correct the trajectories and modes (speeds) of tool movement when they process parts in conditions of uncertainty and variability of the working environment.

With the help of the created new software, dynamic control accuracy using robotic tools of the order of 0.5 mm has already been achieved, even when performing power operations. However, for the implementation of many ultraprecise work operations, accuracy is required within 0.2 - 0.1 mm.

“In the world, this problem hasn’t been solved yet. The problem is the inaccuracy of manufacturing the robots themselves. We have already created a method for eliminating this flaw based on special test movements. On the models, everything works fine. Now we are putting this method into practice. If the results are positive, then we can say that we have made a significant breakthrough in the wider practical use of robots. And if not, we will continue to work to a positive effect. In general, the method works, ”said Vladimir Filaretov, Professor, Doctor of Technical Sciences, Head of the Department of Automation and Control, FEFU School of Engineering.

Using technical vision systems, the machine builds a virtual image of its workspace, independently recognizes each workpiece and determines its exact location in the workspace. For large products, the robot can detect their deformations that occur during fixation. Based on the generated virtual scene, it determines the trajectories of its working tools.

“It should be emphasized that the methods, algorithms, and software that we created are universal. They can be used to control almost any type of robot: industrial, underwater, land mobile, flying and many promising agricultural robotics systems. For their real use, it is necessary to introduce only minor adjustments provided for in the programs and taking into account the specific features of these robots. Our developments, including intelligent control based on virtual reality, make maximum use of the capabilities of modern computing technology, can increase the productivity of the process by not just ten percent, which is good, but by several times at once, maintaining the quality of the products, ”added Vladimir Filaretov.

The new method of intelligent control has already been introduced at the Dalpribor plant (Vladivostok). Now it is being tested and finalized taking into account the latest challenges of the industry. A report on the last of such refinement was also made in Grenoble at the IEEE International Conference on Control, Automation and Diagnosis 2019 (ICAAD’19) and was especially noted by the organizers.

Based on the results of the work, the scientific group led by Professor Vladimir Filaretov consisting of five people applied for the Prize of the Government of the Russian Federation.