Production systems are changing in many aspects on the way to a Factory of the Future, including the level of automation and communication between components. Besides all benefits, this evolution raises the amount, effect and type of anomalies and unforeseen behavior to a new level of complexity. Thus, new detection and mitigation concepts are required. Based on a use-case dealing with a distributed transportation system for production environments, this paper describes the different sources of possible anomalies with the same effect, anomaly detection methods and related mitigation techniques. Depending on the identified anomaly, the FoF should react accordingly, such as fleet or AGV reconfiguration, strong authentication and access control or a deletion of adversarial noises. In this paper, different types of mitigation actions are described that support the fleet in overcoming the effect of the anomaly or preventing them in the future. A concept to select the most appreciate mitigation method is presented, where the detection of the correct source of the anomaly is key. This paper shows how various techniques can work together to gain a holistic view on anomalies in the Factory of the Future for selecting the most appropriate mitigation technique.

In recent years Deep Neural Networks (DNNs) have achieved remarkable results and even showed super-human capabilities in a broad range of domains. This led people to trust in DNNs' classifications and resulting actions even in security-sensitive environments like autonomous driving. Despite their impressive achievements, DNNs are known to be vulnerable to adversarial examples. Such inputs contain small perturbations to intentionally fool the attacked model. In this paper, we present a novel end-to-end framework to detect such attacks during classification without influencing the target model's performance. Inspired by recent research in neuron-coverage guided testing we show that dense layers of DNNs carry security-sensitive information. With a secondary DNN we analyze the activation patterns of the dense layers during classification runtime, which enables effective and real-time detection of adversarial examples. Our prototype implementation successfully detects adversarial examples in image, natural language, and audio processing. Thereby, we cover a variety of target DNNs, including Long Short Term Memory (LSTM) architectures. In addition, to effectively defend against state-of-the-art attacks, our approach generalizes between different sets of adversarial examples. Thus, our method most likely enables us to detect even future, yet unknown attacks. Finally, during white-box adaptive attacks, we show our method cannot be easily bypassed.