Despite decades of technology advances, humans remain a critical component to many manufacturing processes. Humans excel at sensing differences through sight, touch, and even smell depending on the product. We can distinguish subtle flaws, variations or differences and can adjust when faced with unpredictability. Humans are also easily trained. We learn by example and can adapt for new products or requirements.
In comparison, technology brings speed, accuracy and repeatability that we simply can’t match as humans. And despite some human advantages, we also suffer from several limitations. Our eyes can be easily tricked, making us unreliable. Technology makes decisions faster, eliminates bias, and is consistently accurate. However, there are many jobs that are difficult to entirely automate.
In the past decade, a pioneer in the deep learning field suggested that we stop training radiologists, as AI did a better job at spotting potential concerns on patient scans. While this is possibly true, the opinion entirely overlooked the full scope of a radiologist’s job, including treatment plans and interacting with a scared and confused patient. In practice, AI is a great tool to aid, not replace, the radiologist.
It’s similar on the manufacturing floor, where in the United States it’s estimated that 70 percent of manufacturing tasks still require manual intervention or a human decision. In many cases, there are good reasons why a human remains in that position. The cost of automating production for short-run or custom products, for example, makes it more economical to retain human processes. Like the AI-spotting technology for radiologists, there are technology tools that can help make those human decisions consistent, reliable, and trackable. For jobs where full automation is not required, AI-based tools help make human decisions consistent and reliable.
Adding automated decision-support to manual processes helps speed inspection rates, improve end-to-end quality, and provides more qualitative product evaluation to ensure manufacturing are repeatable and traceable. New “app-based” systems combining machine vision and AI capabilities with cameras and edge processing make it easier and more cost effective for manufacturers to add decision-support for incoming, in-progress and finished goods. The scalable suite of apps can be easily customized for unique manufacturing requirements.
Visual Inspection
Integrating machine vision and AI capabilities, camera-based visual inspection can reduce human error and ensure consistent and reliable product quality decisions. In this application, a machine vision-based image compare tool initially highlights differences between a “golden reference” and the product under inspection to reduce quality escapes.
As the operator accepts or rejects those identified differences, behind the scenes they are transparently training an AI model, without requiring expensive and time-consuming coding or labelling. With a few inspections, the AI model will begin suggesting a decision for the operator. Using AI, manufacturers can ensure consistent product decisions across inspection stations, shifts, or locations. The app can be customized to notify supervisors, leads hands or managers of product anomalies.
Product Tracker
As a manufacturer adopts visual inspection tools, they can begin streamlining processes and driving analysis-based continuous improvement by saving, storing, and sharing visual images and data from manual processes.
The Product Tracker app captures, stores, and shares images and operator notes of every outgoing visually inspected or manually assembled product for traceability, inventory management, batch tracking, and to speed customer issue resolution. Product images and data can be stored locally or in the cloud and shared with manufacturing planning systems to gather and analyze end-to-end product information.
RMA Reporting
Leveraging the data gathered by the Product Tracker, an RMA Reporting app helps manufacturers avoid costly product returns. The app acquires product images and operator notes for product returns in a digital RMA Report that integrates with a manufacturing resource planning system to help speed trend and root cause analysis and support continuous improvement initiatives. With data from the Product Tracker app, manufacturers can verify the exact state of a product when it left the facility.
Checklist & Work Instructions
Converting error-prone paper-based processes, including assembly and changeover steps, into electronic checklists ensures staff is using the latest instructions to eliminate errors, waste, and delays. The apps can be used to speed training for new employees and processes, with interactive documents, web links, images, and videos that clearly and visually explain process steps. In addition, the Visual Inspection app can be used to compare against instructions and capture images during in-process steps.
Asset & Tool Management
Misplaced or improperly used tools and assets on the shop floor are a significant source or cost and lost productivity. The camera-based Asset & Tool Management app scans a barcode or unique identifier to “check in” and “check out” assets and tools used in production. The app can be customized to create email and text alerts or a dashboard view so employees know where a tool is and who is using it in real-time, and compile reports so managers can optimize the use of assets.
Digital Job Traveler
Creating digital job travelers and work orders instantly provides operators with the latest schedule, requirements, and priorities they need for their exact job. Digitizing the paper process lets a manufacturer gather a complete picture of their operations, including accurate data that provides real-time visibility as production steps occur and analytics and reporting to help identify and avoid errors. Digital records of all completed jobs, for every product, are more easily stored for audits and regulatory requirements.