Supporting Human-Robot Collaboration and Safety with the Proposed Explainable Neuro-Symbolic Reasoning
摘要
In this article, an overview of the innovative architecture of a hybrid AI neuro-symbolic architecture which uses high-resolution deep vision with probabilistic first-order logic for safety monitoring and anomaly detection in Human-Robot Collaboration environment is proposed. Firstly, the problem and the proposed solution and its application to human-robot collaboration scenarios is outlined. Then, the performance of the proposed method for anomaly detection and its conformity to the requirements defined by the end-users in realistic scenarios is discussed.