Robomat GmbH

The Robomat company cluster Our team is united by great enthusiasm for solving innovative tasks. The main focus is image recognition, user-friendly software, optics and mechanical construction of optical applications. We are big enough to solve your specific tasks, yet small enough to take care of you individually. Our team has been working together since 1999, some of them are also active in other companies in addition to their work for Robomat and add their know-how for our projects. Robomat is an engineering think tank that has its roots in the 1980s. PC software before the PC was invented. We built a computer from the first graphics-capable computer boards and developed a CAD program for precast concrete plants on that hardware. When the first PC hit the market ... ... we were the first to offer a lean CAD software for precast plants executable on a PC. - That brought us to market leadership. Industry 4.0 in the 90s. The development of a master computer and the use of a data connection to the machines made it possible to automatically manufacture prefabricated parts for individually planned houses: series size 1. Formwork robot CAD controlled. Outlines of the prefabricated parts drawn in CAD control a robot. Depending on the individual floor plan it positions side formworks and fixes them with magnets. In order to be more independent from the construction industry we developed a microprocessor-controlled flextime recording device. The entire project was sold. 1996 World's first wall building robot. ... that builds entire brick walls at the factory. The CAD program calculates the position of each individual brick. For the first time we are using image recognition for quality control. Broken bricks are detected and eliminated. 2009 Image recognition for robot control. We solve a first demanding image recognition task: parts scattered on a conveyor belt are scanned with image recognition and parts are selected for picking. Only these parts are chosen which are positioned in such a way that they can be picked up without the gripper colliding with the neighboring parts. The trajectory from the current gripping position is calculated in real time for each part. A highly flexible robot is controlled with the data. The advantage of this powerful robot application is, that 6 pieces of differently shaped parts can be fed in by just one gripper AND if the shape of the parts changes, only the image recognition data have to be adapted for a new job. A further application of image recognition for quality control and automatic error logging. The impetus for this was a tender from BMW for a portable device to test CF fabrics during production.