Fraunhofer Institute for Material and Beam Technology IWS
The real-time processing and data management workgroup specializes in the development of customized complete solutions ranging from real-time capable controls to cloud solutions. Modern industrial processes require increasingly complex controls to assist users in their increasingly demanding work. In parallel to established control and regulation tasks, additional functions such as machine learning, communication, self-optimization and self-diagnosis must be integrated. The keyword is "Industry 4.0".
With the help of the universal control platform developed at the IWS, application-specific solutions can be realized quickly and easily. As with cell phones and tablets, a compact single board computer ("System-on-Chip") acts as the core. These combine a wide range of resources from multi-core processors to interfaces and special circuits to programmable logic (FPGA) and can implement different requirements on a single platform.
In addition to customer-specific sensor and actuator interfaces with hard real-time requirements, the control platform offers extensive administration and monitoring functions up to cloud interfaces through the Linux operating system. Furthermore, additional functionalities such as self-optimization, self-configuration, self-diagnosis can be realized to support or assist users in their increasingly complex work.
Currently, IWS solutions are successfully used, for example, to control complex micro-physiological systems (MPS) and laser material processing operations.
Range of Services
- Development of customized solutions for data acquisition, analysis and control with hard real-time requirements (hardware and software)
- Development of customized sensor, actuator, data and user interfaces
- Support of different standards (SILA-2, CAN, EtherCAT...)
- Development of IP cores and data processing pipelines for hardware acceleration of application specific functions
- IP cores for Camera Link frame grabber, scanner control (XY2-100), and others
- FPGA-based Neural Network Inference