Artificial Intelligence meets motion control: Taking technology a step further
At Wittmann Battenfeld Deutschland, General Manager Michael Tolz is already thinking about the automation of tomorrow: "We always want to make our solutions a bit more intelligent." That's why the company is combining motion control, augmented reality and artificial intelligence in a single controller in a new robot single application.
Wittmann Battenfeld's handling systems are used to grip insert parts for injection moulding machines and then reinsert them into the machine in the correct position. The corresponding object must be moved and rotated freely, so that the handling system has a total of five axes - three for movement in space, and two more to be able to rotate the inserts in any direction. What sounds complicated is actually is an everyday task for humans: "When we grasp and place an object, we don't think about it at all. We rely on our eye-hand coordination, which is controlled by the brain," explains Michael Tolz. As managing director, one of his tasks is to map this everyday human ability in a highly automated process: "And when you look at it on a technical level, you only realise how complex the whole story is."
To make matters worse, the handling systems have to complete this complex task with very fast cycle times so that the entire system works as productively as possible. "But it becomes especially challenging when the user wants to handle different parts in one handling system. Then the system must not only be fast, but also flexible and easy to adapt."
Wittmann Battenfeld Deutschland GmbH
Wittmann Battenfeld manufactures machines and automation systems for the plastics industry and supports users around the world with complete solutions from a single source. The company is one of the leading suppliers of handling systems for injection molding machines. Well thought-out detailed solutions and high-quality components enable users to achieve maximum availability, the shortest cycles and long maintenance intervals.
The task: Flexible gripping and insertion of a wide variety of inserts
Together with Siemens, Wittmann Battenfeld has developed an innovative machine concept for precisely such applications, combining Wittmann Battenfeld's proven mechanics with a technology-oriented control system from Siemens: With the Anyfeeder, the fed parts are first transported from the bunker to a vibrating table, where they are separated from each other by the mechanical movement. A camera above the table then detects the inserted parts and transmits the data to the control system, which recognises them and then positions the gripper so that it grips and deposits the correct ones.
The solution: A lean yet versatile automation system
"In order to make the process as flexible as possible, we opted for a Siemens solution for the automation. With SIMATIC, we have the possibility to map all functions in one controller - motion control, image recognition, automation," Tolz explains further. The core of this is the path interpolation for the movement of gripping and depositing on the workpiece carrier. Wittmann Battenfeld implemented this function with the help of the Handling Standard Application from Siemens, which supports the engineering with modules for motion control tasks and visualisation. Part of the package is also a trace function with which the developers at Wittmann Battenfeld can track the movement of the gripper on the model in 3D.
We always want to make our systems a bit more intelligent in terms of the user.Michael Tolz, General Manager Wittmann Battenfeld Deutschland GmbH
The next step: In-house intelligence for image recognition
Michael Tolz has thought one step further with the Anyfeeder: "We always want to make our systems a bit more intelligent to make operation easier for the customer. For this reason, we have used a SIMATIC Artificial Intelligence module in the Anyfeeder." The SIMATIC S7-1500 TM NPU module works with a special AI-capable processor unit, which features efficient neuronal network processing, among other things. The latter is trained for object recognition using sample objects and is then able to automatically recognise the correct objects and their position on the vibrating table. The AI module then simply passes this data on to the SIMATIC S7-1500 T CPU via the backplane bus, which then controls the gripper accordingly. This eliminates the need for an interface to an additional external unit. The result is a very lean architecture that pays off especially during commissioning, says Tolz: "Since we only have one controller and thus only one machine program and one engineering environment, we can carry out the complete commissioning on site with only one specialist. And for our users, this technology brings significantly more flexibility to their processes - they can change the type of inserts and the module simply learns the new geometry. That's already pretty close to hand-eye coordination as we know it," Tolz explains enthusiastically.
The possibilities of these innovative technologies are far from exhausted - especially when it comes to intelligent operating and service concepts.Michael Tolz, General Manager Wittmann Battenfeld Deutschland GmbH
A virtual environment for operation and service
However, Wittmann Battenfeld does not only use artificial intelligence in the Anyfeeder, says Tolz: "We assume that the possibilities of innovative technologies are far from exhausted. That's why we definitely also want to integrate intelligent operating concepts, especially intelligent service concepts, in the interests of the customer." With the Anyfeeder, users can now view production- and machine-specific data in a virtual environment via augmented reality (AR) during runtime. The data is displayed transparently in the virtual 3D model of the plant and superimposed on the real image of the machine and can be played back on a tablet or even AR glasses. "This makes everyday work easier for the machine operators and service personnel: the user has all the information on site - he does not have to get circuit diagrams, for example, but sees everything he needs directly on the machine and can react immediately. Alternatively, they can use the interface to visualise parts of the system that are difficult to access, in order to locate a fault more quickly, for example. On the one hand, this increases the acceptance of the technologies among the machine operators and service staff because we offer them a real benefit. On the other hand, we assume that this will extremely reduce plant downtimes and significantly increase availability. We already see great potential, not only in plastics automation, but also for many other applications."