SplitCode: Voronoi-based error exaggeration for authentication of manufactured parts

Abstract

Counterfeiting in manufacturing is a crucial problem that has the potential to cause economic losses to both small and large businesses, including the aerospace, automotive, and medical industry. Existing techniques for preventing counterfeiting are based on external modification of the manufacturing process which incurs extra cost, and limits their use in everyday applications. In our work, we take advantage of the inherent characteristics of the machine configuration (machine and process parameters) to identify whether a given part is manufactured by a certain machine or class of machines. Each machine configuration has a unique coordinate error distribution which is indicative of its precision and bias. The overarching idea is to differentiate between error distributions of two machine configurations in order to determine whether the positional errors in a given part come from the same distribution as that of the machine configuration. To be able to differentiate between two machine configurations robustly, we propose a novel topological transformation technique based on the principle of Voronoi tessellation that exaggerates the difference between their error distributions. We present a methodology for authentication of machined parts and validate it numerically and experimentally through the example of additive manufacturing. This research work also offers various opportunities of further exploration in terms of part design, algorithm of SplitCode, imaging and post processing methods and statistical variations.