The Future of Manufacturing: Prototype Robot Solves Problems without Programming

With the help of artificial intelligence, researchers at Siemens have developed a two-armed robot that can manufacture products without having to be programmed. In a glimpse of the future of automated production, the robot’s arms autonomously divide tasks and work together as one.

 

by Sandra Zistl

A click is heard as one hand snaps a gray part into place on a rail. The hand withdraws and grasps another component, this time passing it to a second hand to achieve the best possible positioning, as the two extremities coordinate their movement to assemble part of a control cabinet. Part of a two-armed robot, the collaborative activity of the hands was recently demonstrated at Siemens Corporate Technology (CT) in Munich, the company’s global research unit.  The robotic system is nothing less than a crucial element in the future of manufacturing — a future in which entire factories will control themselves.

Making Batch Size 1 Economical

To some extent, this is already possible in mass production, as demonstrated, for example, by Siemens’ showcase plant in Amberg, Germany. The factory produces Simatic programmable logic controls — with a 75 percent degree of automation and 99.99885 percent quality. However, these parts are manufactured in large batches. Each year, 12 million Simatic controls are shipped to over 60,000 customers around the world. As a result, the future has already become a reality here for high production volumes. The ability to perform tasks autonomously – rather than automatically – is exactly what’s needed for manufacturers of smaller batches and those who produce many different product variants in response to growing demand for customized products.  Conventional automation hasn’t yet been profitable at this level, which is sometimes referred to as “batch size 1.”

 

A team of Siemens Corporate Technology headed by Kai Wurm and Georg von Wichert, who research autonomous systems at Siemens, managed to solve this problem. “Our two-armed smart prototype illustrates that economical batch size 1 production is possible,” says Wurm. “In the future, robots will no longer have to be expensively programmed in a time-consuming manner with pages of code that provide them with a fixed procedure for assembling parts. We will only have to specify the task and the system will then automatically translate these specifications into a program.”

Transitioning to Semantic Information

 “We simply tell the robot to attach a specific component to the mounting rail,” says Wurm. “And that’s exactly what it does.” On a small scale, this task describes what batch size 1 is all about. It involves manufacturing or assembling a product in a wide variety of variants that contain different components. The robot gets the information on how to manufacture a product from an associated software model. Although this CAD/CAM (Computer Aided Design and Manufacturing) model is incomprehensible for conventional robots, the new prototype can understand such models. In a sense, it is as if the robot can understand different languages, thus eliminating the need to program its movements and processes.

 

To do this, the prototype successively divides tasks, such as the general command “assemble,” from the software construction plan into doable units, such as “pick” and “hand over” until it finally moves an arm or opens its grippers. The robot itself also decides which task each arm should perform. To make this possible, the developers have enabled the prototype to raise information from the product development software to a semantic level.  

“Product parts and process information are semantically converted into ontologies and knowledge graphs,” says Wurm. “This makes implicit information explicit. Until now the things that people simply know from experience when they are told to snap component X onto rail Y have had to be taught to robots in the form of code. However, our prototype analyzes the problem by itself and finds a corresponding solution.”

 

In the case of Siemens’ prototype demonstrator, one can witness this process in a vastly simplified form on a monitor to the right of the robot arms. The monitor displays two rows of color tiles, each of which bears words such as “assemble” (left-hand column) and “pick” (right-hand column). These tiles gradually move upward in a manner similar to scrolling down a long webpage. The tiles depict each assembly step. On the monitor to the left, the demonstrator shows the information that the robot arms receive at the beginning of a production process. This information consists of a 3D depiction of the surrounding area and the objects it contains. Above the demonstrator are two more screens that show what the robot arms are currently seeing through their integrated cameras.

Toward Self-Correcting Systems

Siemens Corporate Technologies’ prototype system can also correct faults without having to be told beforehand that this is an option. If a part slips, for example, one of its arms will find the part as long as it is within its camera’s field of vision. The arm will then pick up the component and adjust all of its subsequent movements so that it can still install it correctly. And if the component needs to be snapped into place on the other side of an assembly, the arm will hand the component to its counterpart.  These groundbreaking developments are part of the Company Core Technology (CCT) Future of Automation program. CCTs enable Siemens to focus on crucial fields of innovation such as digital twins, artificial intelligence, and additive manufacturing.

 

Naturally, assembling control cabinets is just the beginning. Siemens developers envision self-organizing production facilities that responds to autonomously changing production requirements, continuously optimize their operations, and are populated by robots that assist one another. Such facilities would be a revolutionary step – essentially systems that feed themselves with design data, corrects faults, and calculates all movements and actions on their own. “There are many other researchers who are trying to solve this problem. But there is nothing comparable to what we have developed on the market yet,” says Wurm.

2017-12-11

Sandra Zistl

Subscribe to our Newsletter

Stay up to date at all times: everything you need to know about electrification, automation, and digitalization.