Combining Measurement Uncertainties with the Probabilistic Robustness for Safety Evaluation of Robot Systems

In this paper, we present a method to engage measurement uncertainties with the probabilistic robustness to one system uncertainty measure. Providing a metric indicating the potential occurrence of dangerous situations is highly essential for safety-critical robot applications. Due to the difficulty of finding a quantifiable, unambiguous representation however, such a metric has not been derived to date. In case of sensory devices, measurement uncertainties are usually provided by manufacturer specifications. Apart from that, several contributions demonstrate that the accuracy of neural networks is verifiable via the robustness. However, state-of-the-art literature is mainly concerned with theoretical investigations such that scarce attention has been devoted to the transfer of the robustness to real-world applications. To fill this gap, we show how the probabilistic robustness can be made useful for evaluating quantitative safety limits. Our key idea is to exploit the analogy between measurement uncertainties and the probabilistic robustness: While measurement uncertainties reflect possible shifts due to technical limitations, the robustness refers to the tolerated amount of distortions in the input data for an unaltered output. Inspired by this analogy, we combine both measures to quantify the system uncertainty online. We validate our method in different settings under real-world conditions. Our findings exemplify that incorporating the novel uncertainty metric effectively prevents the rate of dangerous situations in Human-Robot Collaboration.

I. Introduction

Assuring safety in robot applications, specifically in human-involved systems, is of central importance. Often, the development of such systems is bound to finding a quantifiable measure that reflects the criticality of situations at system run-time. For instance, a robot operating in an environment with humans must reduce its speed once the detection performance of humans or objects falls under a cer-tain threshold. In particular for industrial robot applications, the compliance with safety limits (e. g., ISO 13849) requires the knowledge of the probability for dangerous failures per hour (PFDH). However, an appropriate metric that indicates the potential occurrence of this probability online and can directly be used to initiate the adaptation of robot parameters (e. g., speed reduction) is missing to date. In this work, we present a novel system uncertainty quantification technique by engaging measurement uncertainties with the probabilistic robustness of neural networks (NNs). Basically, we argue that undesired distortions in data due to technical limits, environmental disturbances or a lack of system knowledge are one dominant source for unexpected system behavior. Fig. 1. Overview: To derive a system uncertainty metric, we draw analogies between measurement uncertainties and the probabilistic robustness of neural networks. Building upon state-of-the-art literature, we assign practical meanings to all parameters.