No agriculture, no stability, no work, no strength. As an industry with strong hematopoietic function, processing and manufacturing industry plays an important role in the sustainable prosperity of economy and social stability.

With the development of industry, human beings have greater ability to transform nature and obtain resources. The products they produce are directly or indirectly used in people's consumption, which has greatly improved people's living standards. It can be said that since the first industrial revolution, industry has determined human survival and development in a certain sense.

However, the rise and fall of industry. In recent years, due to the hollowing out of industries in developed countries and the low value of industries in developing countries, the dilemma of processing and manufacturing industry has emerged. A large number of workers in developed countries are unemployed and have trade deficits, and the profits and environment of developing countries are deteriorating. A large number of manufacturing enterprises are facing survival crisis, and the transformation and upgrading of manufacturing enterprises in digitization, networking and intelligence is imminent.

At the same time, with the rapid development of artificial intelligence technology and its wide application in the field of consumption and circulation, more and more manufacturing enterprises and artificial intelligence enterprises pay attention to "artificial intelligence + manufacturing". However, at present, "Ai + manufacturing" still has the problem of insufficient power, and the road of AI in manufacturing industry is still long.

AI manufacturing dilemma still exists

The manufacturing industry enabled by artificial intelligence technology has great potential. The combination of artificial intelligence and related technologies can optimize the efficiency of various process links in the manufacturing industry, collect various production data through the industrial Internet of things, and then provide suggestions or even independent optimization after processing with the help of deep learning algorithm.

From the application scenario of artificial intelligence in manufacturing industry, it mainly includes product intelligent R & D and design, the use of artificial intelligence in manufacturing and management processes to improve product quality and production efficiency, and the intellectualization of supply chain.

In product R & D, design and manufacturing, artificial intelligence can not only explore various possible design solutions and carry out intelligent generative product design by using algorithms according to the established goals and constraints, but also integrate and commercialize the achievements of artificial intelligence technology to produce a new generation of intelligent products such as smart phones, industrial robots, service robots, autonomous vehicle and UAVs.

For production and manufacturing, the embedding of artificial intelligence into production and manufacturing links will make the machine smarter, no longer only perform monotonous mechanical tasks, but can operate independently in more complex situations, so as to comprehensively improve production efficiency.

In intelligent supply chain, demand forecasting is the key theme of applying artificial intelligence in the field of supply value management. By better predicting demand changes, the company can effectively adjust production plans and improve plant utilization. In addition, the intelligent handling robot will realize the independent optimization of warehousing, greatly improve the efficiency of warehousing and sorting, and reduce labor costs.

However, whether it is intelligent R & D and design, production and manufacturing, or intelligent supply chain, manufacturing digitization is the basis of artificial intelligence + manufacturing. However, the informatization level of China's manufacturing industry is uneven, and the manufacturing industry chain is far more complex than other industries, which emphasizes the understanding of the industry background of the enablers. All these have resulted in higher threshold and greater difficulty in Al empowerment of manufacturing industry compared with other industries.

Manufacturing industry is a huge industry, which is complex and fragmented. In the same plant, there are often several production equipment from different manufacturers. These equipment often adopt their own technology and data standards, which can not be directly connected and interactive with each other. Different factories and even different manufacturing enterprises have even greater differences. Such differences make it difficult for traditional manufacturing informatization and limited efficiency improvement.

Although artificial intelligence technology has been applied in some links and processes of manufacturing industry to a certain extent, the overall penetration rate is still at a low level. According to the calculation of China Academy of communications and communications, the proportion of China's industrial digital economy in 2018 was only 18.3%, less than 20%. Under the background of the low overall digital level of manufacturing industry, the penetration rate of artificial intelligence technology in the digital economy of manufacturing industry is obviously lower.

In addition, at this stage, the value of artificial intelligence is still difficult to be accurately measured, and some enterprises, especially small and medium-sized enterprises, lack the power to apply artificial intelligence. The reason is that the application of some technologies in the field of artificial intelligence often aims to improve the brand and increase product empowerment, so as to improve the profit margin or reduce the internal operation cost. However, due to the small volume of small and medium-sized enterprises, they often take survival as the minimum goal. If they need to open their market, most choose to start from increasing revenue and reducing expenditure.

In other words, small and medium-sized manufacturing enterprises focus on efficiency when building intelligent systems, but they get efficiency at the cost of a lot of costs. In other words, there is no real balance between efficiency and cost.

In addition to the profit eating behavior of small and medium-sized enterprises, even if the large enterprises standing in the first echelon are not clear about the application path of artificial intelligence in some subdivided industries, it is difficult to accurately calculate the application risks, benefits and costs, and can not give practical solutions in a short time. Coupled with years of overcapacity, despite the huge amount of data, it will take a long time to realize intelligence.

Human is the core of intelligent manufacturing

There are still practical differences between the intelligent process of manufacturing industry and the automation of manufacturing industry in the past. Intelligence is not equal to automation, let alone unmanned. How to move towards intelligence is related to solving the current AI manufacturing dilemma and the real landing of the transformation and upgrading of processing and manufacturing industry.

Automation pursues the automatic production of machines, and its essence is "machine replacement", emphasizing large-scale machine production; "Intellectualization" pursues the flexible production of machines, and its essence is "man-machine cooperation", emphasizing that machines can independently cooperate with the changes of factors and people's work.

It can be seen that intelligence must not be equal to unmanned. In the process of promoting a large number of intelligent manufacturing, only by promoting the change of decision-making and thinking through the integration of machines and people, can people's working ability and direction be expanded and the empowerment of machines be maximized.

Therefore, what AI + manufacturing pursues is not a simple "machine replacement", but to pull the extremely refined and even alienated workers' assembly line work since the industrial revolution back to the "people-oriented" organization mode, so that machines can undertake more simple, repetitive and even dangerous work, while people can undertake more management and creative work.

Obviously, if we want to realize the intelligent processing and manufacturing of man-machine integration, we must go through the process from man to machine. Only when the machine integrates more intelligence can it expand more capabilities. The application of industrial robot is an important symbol of this stage. As the perfect combination of industrialization and informatization, industrial robot has opened the connection between a single production equipment and the whole production network with its natural digital characteristics, and then supported the application scenario of the fourth industrial revolution.

If the development of the Internet in the past two decades has connected everyone in the intelligent era, the development of industrial intelligence in the next two decades will connect every industrial robot, thus bringing comprehensive innovation in production efficiency and even production mode.

However, in the process of realizing from human to machine, industrial robot also needs to have the attribute of interacting with human in complex and atypical environment. Only by being flexible and convenient can we meet the development conditions of man-machine integration and make a comprehensive deployment for the intellectualization of manufacturing industry. In addition, the deployment of machines should also be scalable, that is, more intelligent platforms need to be equipped to expand the application scenarios of industrial manufacturing.