int(1076)

Nobleo’s definition

for Artificial Intelligence

Nobleo’s definition for Artificial Intelligence

Thanks to Moore’s law, these days human mankind has access to powerful computing platforms. This enabled massive innovation in the field of Artificial Intelligence in past years. It is about collection and -storage of huge amounts of data, and the related processing of that data with trained algorithms into information. That information on its turn is used to create intelligent machines or processes in use by humans.

Around Artificial Intelligence and Machine Learning there is quite some buzz, and that is the reason that at Nobleo we prefer to speak about Intelligent Solutions.

An Intelligent Solution is a machine or a process that enables automated decision making based on (intensive) data collection and interpretation to continuously optimize system performance. To obtain this data, an intelligent system contains either sensors and/or uses existing signals/data sources. Such a system can have a function at design time or manufacturing time or when in use by an end-user.

Example of a design time usage is a tool that is used to develop products. These days the layouts of complex computer chips are made by design tools which are based on trained algorithms, and can execute the complex and iterative component placement and connection process way faster than a human can do.

Example of a manufacturing time usage is the interaction between a manufacturing or inspection process and a product being created or handled by a machine executing that process. Imagine a printer executing a print job, and corrections in material transport or environmental conditions are done while printing, based on trained algorithms that take action because of varying and measured conditions inside the printer. By working this way, the creation of products that are out of specification and that have to be reworked or even scrapped is reduced.

Example of an end user usage is an operator in a factory applying an inspection tool to inspect products for failures. Imagine the manufacturing of bare printed circuit boards, which have to be inspected for interrupted tracks, short circuits and other potential failures. As this type of failures can occur everywhere in the product in any combination and failure rate, the best approach for this kind of applications is to apply trained algorithms.

Another dimension of Intelligent Solutions as we define them at Nobleo, is that such solutions should solve actual problems for users. And this automatically means that the voice of the end user should be taken into consideration, and that end user application knowledge should be incorporated in the intelligent solution. Due to our collective and multiyear experience in mechatronics equipment development and in the creation of autonomous systems, we have in most cases this domain knowledge already available in house. And in those projects where we lack the application knowledge, we take care that we organize a true co-development with our customers to accommodate for this.

November 2021