Toward Safe Human-Robot Interaction: More Efficient Safety Monitoring Technology For Robotic Arms

Presently, advanced humanoid robotic arms have the ability to work independently, and although most current research can make the robotic arm react appropriately and autonomously to the changing environment, it cannot completely prevent the robotic arm from reacting incorrectly due to a variety of uncontrollable factors; therefore, it is necessary to detect the overall posture of the operating robotic arm. In this paper, we propose a more effective vision fusion-based robot safety monitoring method that can detect the robot arm’s posture change in real time, thereby significantly enhancing the safety of the robot arm at work. The visual fusion method utilized in this paper enables faster and more stable skeleton visualization tracking of the robot arm’s motion, and at the same time, more accurate skeleton information extraction of the robot arm in three-dimensional space, which increases the accuracy of detecting the bending angle of each joint under each posture of the robot arm. The research experiment demonstrates that the method proposed in this paper has a positive effect on monitoring the robotic arm in real time.

I. Introduction

With the progression of time, the traditional manual production method is being displaced by the human-machine collaboration production method [1], or even a “machine replacing human’’ mode. This indicates that the production process will involve less human involvement. “Machine replacing human’’ has become popular, and collaborative robots can significantly increase productivity to satisfy the high precision, customization, and adaptability requirements of production. This has resulted in a reshaping of production methods and industry patterns, with the use of collaborative robots becoming more prevalent across a variety of industries and the demand for collaborative robots in industry increasing. Likewise, safety monitoring of collaborative robots has become increasingly crucial. On the one hand, it aids in the prevention of potential safety hazards, and on the other, it effectively reduces losses caused by collaborative robot failures. Consequently, safety monitoring of collaborative robotics is currently a popular research topic.