Russian Researchers Teach Industrial Robots to Understand Operator Gestures
The system enables contactless control of robotic manipulators using computer vision and neural networks.
Researchers at Donskoy gosudarstvennyy tekhnicheskiy universitet (Don State Technical University) and MGTU Stankin (Stankin Moscow State Technological University) have developed domestic software that allows industrial robots to be controlled through hand gestures.
The system relies on computer vision and neural network algorithms, combining input from a standard camera and a depth sensor. The camera captures visual data, while the depth sensor measures distance to objects. Together, they compensate for each other’s limitations, including in low-light conditions. The software is trained to recognize 10 gesture commands and transmit them to a robotic manipulator controller.
The technology can be used to automate tasks such as loading, assembly, painting, laser processing, and marking. The project was supported by the Prioritet 2030 (Priority 2030 program) initiative and a grant from the Russian Science Foundation.
