An efficient dataflow-oriented fuzzy library

Fuzzy logic is often used in intelligent systems in general and especially in learning domains. On this occasion, always returning problem formulations are to be solved like an arithmetic- and memory-efficient implementation by fuzzy operations. We present a software library which aims to an efficient, extendable and comfortable implementation for embedded mechatronic systems. The data-flow oriented concept was explained. It enables a modular and extensible system design where the communication path and also the behavior parameters are exchangeable. We summarize the implementation issues and give an overview of the special type of operations on a memory and processing restricted controller. The evaluation is done with a lab robot system having an omnidirectional drive. We implemented the controller in a commercial engineering design environment and compared the behavior of our library with a common library which comes with the commercial solution. The results show that our implementation not only works but also shows good performance on the robot drive. The conclusion shows the portability of the library to an application example of movement analysis by fiber optical sensors.