Introducing InnerVoice

InnerVoice is a real-time, in-process monitoring hardware and software platform for manufacturing quality assurance. InnerVoice combines AI and machine learning into a powerful, general-purpose, data-driven process monitoring system that enhances the productivity and reliability of your manufacturing process. InnerVoice condenses the raw manufacturing process data into a unique digital fingerprint on a part-by-part basis and immediately alerts the manufacturer to process outliers when novel or anomalous behavior is encountered. InnerVoice catches flaws when and where they occur and before they become defects, thereby preventing bad parts from being delivered to the customer.

InnerVoice is the result of 30 years of research, 10 years of development, and 5 years of beta testing. It was designed from the ground up to be a pluggable architecture. This design approach allows for a "process-agnostic" monitoring framework where one or more plug-ins provide a customized feature set that is specifically tailored to the customer's data, process, machine interface, and reporting needs.

InnerVoice provides four main capabilities:

Visualization: First and foremost, InnerVoice is a visualization tool that helps manufacturing engineers view complex and often times noisy data. In manufacturing, plotting and visualizing the raw process data is an absolute requirement for understanding and gaining deeper insight into the complex behavior of the manufacturing process. InnerVoice automates and simplifies the tasks of organizing, loading, cropping, filtering, transforming, and visualizing your manufacturing process data.
Exploration: Beyond visualization, InnerVoice provides a rich set of feature extraction and exploration tools that help manufacturing engineers identify relationships and achieve a deeper understanding of their manufacturing process. Manufacturing process data is usually complex, non-linear, and tightly coupled. InnerVoice contains techniques that map the high-dimensional data to lower dimensionality for exploration and allows a manufacturing engineer to draw informed conclusions and make better decisions.
Analysis: At its core, InnerVoice is a real-time, data-driven computational engine that supports signal processing, feature extraction, pattern recognition, and classification. InnerVoice condenses the raw manufacturing process data into a unique digital fingerprint on a part-by-part basis. This fingerprint contains salient details about the machine and process behavior. Using this fingerprint, InnerVoice applies AI and machine learning techniques to inspect the process performance of each part looking for outliers with novel or anomalous behavior. Outliers can then be passed on to a human for additional scrutiny, thereby creating an "inspect for cause" quality assurance methodology.
Communication: InnerVoice can communicate directly with PLCs controlling the manufacturing process or with the manufacturer's IT database. This direct line of communication provides data-driven, actionable intelligence and enables real-time visibility and fine-grained control over the manufacturing process. Flaws are identified, contained, and corrected without further value-added downstream processing.

InnerVoice's pluggable architecture enables it to be customized to accommodate a customer's data, machine interface, manufacturing process, analysis, and communication needs. This modularity prevents unwanted or unnecessary features, simplifies the user interface, reduces the memory footprint, and improves performance. Most importantly, plugins can be developed quickly and cost-effectively and without modification to the main application framework.

In 2016, we released InnerVoice to a small group of manufacturers from three different continents for beta testing. During this five-year period, InnerVoice inspected over 2.5 million parts. With their valuable feedback, we implemented numerous features and enhancements that simplify the integration and improve the capabilities of InnerVoice.

Scrap is inevitable, but delivering defective parts or products is preventable. InnerVoice thoroughly scours your manufacturing process data on a part-by-part basis searching for outliers with novel or anomalous process behavior. Catching flaws when and where they occur enhances productivity, minimizes downtime, and avoids unnecessary and embarrassing customer relationship problems.

Resources D.A. Hartman 04 July 2021 Hits: 12291