Predictive Sensor Analytics for Power Side-Channel Monitoring of Security Applications in Fused Deposition Modeling

Abstract

The advent of Additive Manufacturing (AM) changed the landscape of the manufacturing industry resulting in new possible designs and products. Around the same time, Industry 4.0 got great attention from the industrial community. Industry 4.0 works across systems connecting machines and different sub-systems and collects and analyzes data to make decisions. This type of interconnected system is considered to be the direction in which the manufacturing world is moving. Thus, it is safe to say that AM will be incorporated into Cyber-Physical Systems (CPS) in some capacity. A CPS consists of three components: a cyber component, a physical component, and a combination of these two components. An increased number of interaction points in the CPS leaves the system vulnerable to attacks, either in terms of intellectual property theft or the system can be sabotaged. Traditional cybersecurity protocols may mitigate the attacks to some extent, but they can not handle the physical component of the system, keeping space for some attacks. In such scenarios, side channels like sound, vibration, and heat may provide certain information regarding the process, irrespective of whether the system is compromised or not. This work is aimed at analyzing the power side-channel signal in a Fused Deposition Modeling (FDM) process, a type of AM process. The main similarities and differences between a cyber system and CPS and their different strategies are discussed. A method is designed to predict the power signature based on the CAD design. A comparative study of the predicted and actual power signature can provide certain information about the machine and process, including anomaly or sabotage detection. ii