Anomaly Detection for Dynamic Human-Robot Assembly

Abstract

Human-Robot Collaboration (HRC) requires humans and robots to work on the same product in the same work environment at the same time. Therefore, the robotic system needs to understand human behavior so it can assist the human appropriately. Since the human is an uncertain variable in this system, human action recognition is one of the key challenges when it comes to HRC. To address this problem, we developed an anomaly detection framework for the dynamic assembly of complex products. We used an Long-Short-Term-Memory (LSTM)-based autoencoder to detect anomalies in human behavior and post-process the output to categorize it as a green or red anomaly. A green anomaly represents a deviation from the intended order but a valid assembly sequence. A red anomaly represents an invalid sequence. In both cases, the worker is guided to complete the assembly process. We demonstrate our proposed framework using an appropriate industrial use case.