Endpoint Security
,
Governance & Risk Management
,
Internet of Things Security
Success Hinges on Marrying Programmed Task and Information From Production Settings
Rockwell’s automation efforts have moved away from a purely programmed approach to one that combines programming and self-learning based on specified parameters.
See Also: Live Webinar | Generative AI: Myths, Realities and Practical Use Cases
The Milwaukee, Wisconsin-based industrial automation vendor said programming all the different use cases for an autonomous vehicle and directing how much pressure should be applied to the accelerator or brake based on traffic signals or road conditions simply wouldn’t be feasible, according to Senior Vice President and Chief Technology Officer Cyril Perducat.
Perducat said Rockwell Automation has instead trained autonomous vehicles using millions of images that capture the optimal behavior of human drivers under various conditions. Since Rockwell cannot predict or anticipate ever single situation an autonomous vehicle will encounter, the vehicle must marry programmed tasks with new information learned in real time while in the field (see: Rockwell Forges Gen AI Pact With Microsoft, Buys Cyber Firm).
Extending the behaviors or practices of high-performing employees to machines through autonomous operations means the processes of experienced workers can be replicated across an entire organization, he said. For instance, an experienced worker would know how changing weather affects manufacturing processes, and in the future, the whole company can benefit from the collective learning, Perducat said.
Going forward, industrial organizations will need to employ both individuals with domain expertise as well as data scientists to obtain value from all the information that’s being generated, he said. This will help businesses move from a system that’s programmed with all tasks anticipated in advance to one that’s truly autonomous and can broadly improve the quality of production taking place, Perducat said.
Charting a Distinct Path With Real-Time Data
Transitioning to autonomous operations means a device must be able to adapt to what happens without having a predefined path, Perducat said. The organizations must be able to adjust the parameters in real time to achieve the optimal production outcome, look at dependencies between multiple factories or across the enterprise, and combine all the ingested data to determine the best possible protections.
The key outcome expected from industrial production environments has long been the right product at the right time, according to Perducat. With the rise of autonomous operations, Perducat said, companies increasingly want the agility to manufacture dozens of different products on a single line rather than just one single product and the ability to evolve and manufacture products that haven’t yet been invented.
Autonomous operations also can help with reducing energy consumption and emissions associated with the manufacturing process, and Perducat said when and how much energy is being used can also become a parameter. Automation can help with creating business optimization scenarios that humans couldn’t optimize on their own based on the volume of parameters and variables involved, he said.
Summarizing, Perducat said autonomous operations can help with resolving complex optimization scenarios and ensure data is being used to the maximum extent possible to simplify matters for people operating production lines. He said 80% of use cases can be automatically addressed and parameters can be adapted continuously, enabling humans to focus their energy and knowledge on different use cases.