Uncertainty Estimation for Safe Human-Robot Collaboration Using Conservation Measures

We present an online and data-driven uncertainty quantification method to enable the development of safe human-robot collaboration applications. Safety risk assessment systems are strongly correlated with accuracy measurements: Distinctive parameters often not directly accessible via known models must therefore be measured. However, measurements generally suffer from uncertainties due limited performance sensors, even unknown environmental disturbances, or humans. In this work, we quantify these measurement by making use conservation measures which quantitative, system specific properties that constant over time, space, other state space dimensions. The key idea our lies in immediate data evaluation incoming during run-time referring equations. particular, estimate violations a-priori known, domain consider them as consequence uncertainties. validate on a case context collaboration, thereby highlighting importance contribution for successful robot under real-world conditions, e.g., industrial environments. addition, show how obtained values can mapped arbitrary safety limits (e.g., ISO 13849) allows monitor compliance standards run-time.