Automation of industrial artificial intelligence
time：2022-03-21
browse：263
Parallel to the development based on quality control theory is automatic manufacturing technology. From 1940s to 1950s, CNC machine tools were invented and officially commercialized, which played a subversive role in the machining industry and greatly improved the production efficiency and machining accuracy.
The development of microprocessors in the 1970s further reduced the cost of CNC machine tools and made them more widely used. In addition to CNC machine tools, industrial robots have also developed in almost the same period. In 1956, American inventor George dewall and physicist Joseph ingelberg invented the world's first industrial robot, unimate.
With the gradual popularity of NC machine tool technology and integrated circuit, faster and simpler industrial robots can be designed and manufactured. The manpower of industrial manufacturing system is gradually reduced and the degree of automation is higher and higher. However, the field of automation and control science in industry has always been self-contained, and the concern is functionality.
Whether or not a certain machine has the consciousness to design or not, that is, whether or not a certain series of actions can be completed by the machine itself, does not care about whether or not the machine has the consciousness to design or not. Today, control theory has the trend of combining with artificial intelligence. For example, automatic driving technology is an automatic product integrating artificial intelligence.
The premise of the concept of "industrial artificial intelligence" is that artificial intelligence technology begins to have the conditions for engineering. If we understand the development history of artificial intelligence, we will find that the mainstream field of artificial intelligence has been divided into two fields: general AI and narrow AI, and the emphasis on the development level of machine ability comparable to or even beyond human general artificial intelligence determines people's attention to the technology in this field of artificial intelligence.
The cost of computing and processing unit, the richness, durability and cost of sensors and other hardware infrastructure are the prerequisites for the implementation of artificial intelligence technology. With the increasing maturity of industrial equipment automation, the deployment of a large number of sensors and the improvement of computing power, concepts such as industrial Internet, industrial Internet of things and industry 4.0 begin to emerge one after another. The reason is that the time is ripe. No matter what the future industry will be called, people are ready to embrace change.
However, there is a big challenge to apply industrial artificial intelligence directly to the industrial field, that is, the machine learning model can not train itself with the experience and knowledge in the minds of experts. At the same time, throughout the development history of industrial intelligence, it is only a technical means for enterprises to make industrial systems intelligent.
Even today, improving efficiency, reducing costs and ensuring quality remain the same theme. Therefore, the closer way to landing is to apply narrow AI technology in the industrial field to solve the problems with exact definitions and clear boundaries. AI scientists should combine the domain knowledge and use the data generated by the equipment to train industrial intelligent applications with specific functions and realize the purpose of intellectualization in a specific problem domain.
