A new intelligent control and collision detection system of robot arm based on machine vision

This paper presents a novel collision point detection model integrated into a robotic arm control system, utilizing a six-dimensional force sensor and machine vision technology. Unlike traditional applications, this model employs collision detection methodologies to enhance system responsiveness and categorizes collisions based on impact points and external forces, enabling accurate analysis of different collision types, including single-point and multi-point impacts. We detail the design and mathematical modeling of an orthogonal parallel six-dimensional force sensor, highlighting its components and operational principles. To validate the feasibility and accuracy of the proposed model, we constructed a three-degree-of-freedom robotic arm for simulation experiments. The results include a comprehensive error analysis of experimental and calculated forces, demonstrating that while experimental values exhibit greater fluctuations, the formula-based calculations provide a consistent trend that reflects the relative errors within the control system. Overall, this study advances the understanding of collision detection in robotic applications, offering a foundation for future research to enhance accuracy and reliability in dynamic environments.