Multidimensional Impact Characterization in Collaborative Robotics: A Gazebo-Based Collision Testing System

This paper developed a collaborative robot collision testing system incorporating a human model and multi-modal sensors, accompanied by corresponding testing methodologies, to verify compliance with ISO/TS 15066 safety specifications. The comprehensive system comprises three modules: a human model testing platform with multi-modal sensing capabilities, a collaborative robotic platform, and a data acquisition/processing unit. The human model testing platform, featuring human posture simulation functionality, integrates internally embedded three-dimensional force sensors, six-axis force/torque transducers, and inertial measurement units to achieve synchronous acquisition of multidimensional parameters during collision events. By establishing a dynamic collision testing platform, the system quantitatively analyzes critical impact indicators including peak collision force, contact duration, and energy transfer characteristics during human-robot interactions, enabling effective assessment of potential biomechanical injury risks. The research implemented collision process simulations within the Gazebo virtual environment and experimentally validated system efficacy. Experimental results demonstrate the system's capability to accurately capture three-dimensional force vectors, torque distributions, and kinematic acceleration data at collision interfaces, thereby providing multidimensional evaluation criteria for robotic safety performance assessment.