Artemis is a cloud-based data collection hub which uses artificial intelligence to power each of the components in the SmartDFE. The data collected anonymously is used to model different systems and track them over time. The data is put through an advanced artificial intelligence system to acquire knowledge which can be passed on to manufacturers or downloaded back to the SmartDFE to make the system more effective. Once the training has been done in the Cloud, the intelligence can be run locally at the SmartDFE (or other applications). Different levels of help can be provided by Artemis - everything from enhancing automated control to print operator guidance.
Features and benefits
- Data collection from all SmartDFE applications into a central cloud database
- Aggregation of data across applications (SPC, Color, Screening, and Performance)
- Historical and trend analysis
- Interface to MATLAB allowing the use of Big Data and AI analysis
- Training and deployment of Artificial Intelligence (AI) models
- Predictive maintenance scenarios
- Cost analysis and prediction
- Accurate prediction of PC requirements for a new printer
- Prediction of maximum line speed for a specific job
Centralized data collection and aggregation
Artemis provides a standard way to collect application data and store it in the Cloud. Artemis will store the information locally whilst the application is off line and will then upload it when the connection returns. The data from all the applications can be tagged with a printer, media and Ink IDs to allow for aggregation of data from different applications giving a comprehensive model based on multi-application feedback.
Historical and trend analysis
The data can be stored over time allowing the data to be reviewed at a later date. The historical data can be analyzed for trends and patterns highlighting changes that could signal a problem.
Interface to MathWorks® MATLAB
The data in the Cloud can be transferred into MathWorks MATLAB to allow data scientists and engineers to run powerful experiments. The data sets will grow very large and therefore MATLAB's Big Data functions will allow the researcher to still run experiments and allow visualization of the data. The MATLAB AI capabilities enable researchers inside and outside Global Graphics Software to deploy AI and deep learning technologies.
Training and deployment of AI Models
Although the training and creation of the AI models will be done in the Cloud using MATLAB, the resulting models can be deployed locally in the SmartDFE applications. This means that even if the production press is not connected to the internet, the AI models can still add intelligence to the SmartDFE. You can use reference printers in your factory and at friendly customers to generate models used by all customers.
Predictive maintenance scenarios
Using the trend data from the SPC and the other applications we will deliver intelligent predictive maintenance services. Predictive maintenance has proven the most cost effective way to keep printers running 24/7. It is the perfect balance between periodic maintenance, which may replace perfectly good parts, and running to failure, which always seems to happen at the worst time. Predictive maintenance tracks key data and uses models trained on trends to predict parts that are actually entering a failure state before the failure happens.
Cost analysis and prediction
There are many ways to cost jobs; some conservative methods and others can under-predict costs causing potential losses. The Smart Job Cost Estimator (JCE) takes a PDF and uses the same technology as is driving the printer to generate a cost. However what about maintenance costs, human costs, and consumables? You can add this data in manually based on some anecdotal or basic equation. However the information that comes back from Artemis and SPC would allow for richer cost prediction models. These models could be downloaded into future JCEs to allow the sales team to feel more confident about quoting for jobs.
Accurate prediction of PC requirements and maximum line speed
RIPping jobs and driving the head driver electronics directly has huge benefits for the printer manufacturer and the final customer. However the printer vendor has to choose the right PC hardware to drive their printer at full rated speed. The problem of choosing hardware is the main reason Direct Benchmark was created and adding Artemis allows many PCs to be benchmarked and compared. Using these centrally stored benchmarks we can start to predict what PC will be required for a specific job and line speed. In a similar manner, we can download the models for the SPC to allow us to predict the line speed achievable with any specific job. Using this information, a print operator can choose to slow a printer down, streamline a the job before submitting, or reject the job.
Working with you
Here at Global Graphics Software we understand how difficult it can be to get started on your digital print journey, and so our Technical Services team is ready to offer free expert advice and help.
Dedicated experts in RIP and screening technology, principal software engineers and color scientists, all with decades of real-world experience, are ready to guide you through the process of selecting the right software components and then successfully implementing them into your press.
We pride ourselves on offering a personal, friendly service and work with you to ensure the fastest time to market.