Strong AI+ Robot Collaboration Will Work Together To Build A Smart Factory Of The Future

- Oct 22, 2019-

With the development of artificial intelligence, robotics, industrial Internet and other technologies deep into the production line, the degree of intelligence in the manufacturing industry has deepened and challenges have come one after another. At the Intelligent Manufacturing System and Independent Decision Forum of the World Intelligent Manufacturing Conference, experts and scholars from China and Germany have conducted extensive exchange discussions on how to further enhance industrial intelligent production.


After years of promotion and development, robots have become the standard of many enterprise production lines, becoming a representative technology of intelligent manufacturing, and currently facing the challenge of upgrading and transformation. Guo Xin, director of global expansion strategy of Siemens Digital Industry Group, said that the popular tandem robot is a common one-arm robot. It is still in the early stage and cannot be intelligent. "From the perspective of global industrial development history, the current robot is equivalent to industrialization. Early rigid line workers, although highly automated, can only produce a single product." As the company's demand for production intelligence increases, robots are also facing upgrades.


How to transform and upgrade the current industrial robot? Zhang Dan, a member of the Canadian Institute of Engineering, introduced a method of thinking: parallelizing the tandem robots, that is, installing three robot arms on a single spindle to achieve coordinated operation. In contrast, parallel robots are faster, more accurate, more efficient, "especially its accuracy, such as in medical robots, we can achieve surgical operations with an accuracy of 0.5 microns, which can also be used in aviation, electronics, etc. requiring high precision machining. The industry. "There are more robotic arms, and new requirements are put forward for robot technology. "The need to solve the problem of interference between the corner and the multi-arm is to make the cooperation between the three robot arms smarter and prevent each other from interfering with each other. produce."


In fact, robotic collaboration places high demands on the intelligence of the system. Mei Tao, a researcher at Pengcheng Laboratory and an expert in the National Defense Science and Technology Innovation Zone, has been working on the research and practice of underwater robot collaboration. He shared his team's experience in multi-robot collaboration. “Compared with single robots, collaboration can not only make Multiple robots collaborate to perform complex tasks and form a modular robot cluster, which greatly reduces system cost."


According to their practice, to maximize the cooperation of robots, it is necessary to establish an intelligent network system with a high degree of artificial intelligence. "We are now based on the knowledge map, designing an optimized graph search algorithm, and dynamically generating decision-making execution of various tasks. Sequence; based on sequence organization of underwater robots to carry out parallel scheduling of job tasks, complete the technical operation of the entire platform."


If the robot is the body, then AI is the brain, real intelligent manufacturing, AI must be developed. Zhang Jianwei, an academician of the Hamburg Academy of Sciences in Germany, said that the technology currently used in industrial production is weak AI, which is mainly to assist people in production. "Our now-called AI can be read and even written, but it is far from thinking. This involves not only algorithms and chips, but data is also very important. According to Siemens' budget, by 2025, there will be approximately 100 billion devices connected to the Internet of Things worldwide, and the resulting data will be Amazingly, how to use these massive amounts of data will be the key to future manufacturing development.


At present, through the integration of AI, big data and other technologies, Siemens has practiced full digital production in an unmanned bus company. Daily driving time, mileage, battery operation... In Siemens' mindshpere system, not only can the operation of an unmanned bus be recorded, but it can also be traced back to the source of production. Guo Xin, Director of Global Development Strategy, Siemens Digital Industry Group "If there is a problem with overheating of a part, we can not only know it in time, but even trace it to its production design stage, thus fundamentally improving the production design of the entire vehicle."


With the advancement of artificial intelligence, big data, robotics and other technologies, how far can smart manufacturing go? At the forum, Siemens described the future of new energy vehicle production workshops: there is no long production line, only one unit format production station, they are responsible for the production of their own parts, unmanned vehicles between the production stations. All the parts are finally assembled by robot assembly. Experts admit that at present, such a picture is only a good wish, but the future is so probable as technology and industry continue to advance.